Grant  Agreement  number:  FP7-­‐KBBE-­‐266124    Project  acronym:  GREENLAND    Project  title:  Gentle  remediation  of  trace  element-­‐contaminated  land    Funding  Scheme:  Collaborative  project    Period  covered:  from  January  1  2011  to  December  31  2014    Name   of   the   scientific   representative   of   the   project's   co-­‐ordinator:   Dr.   Markus  Puschenreiter,  Universität  für  Bodenkultur  Wien    Tel:  +43  1  47654  3126    Fax:  +43  1  47654  3130    E-­‐mail:  markus.puschenreiter@boku.ac.at    Project  website  address:  www.greenland-­‐project.eu    

PROJECT  FINAL  REPORT  

1.  Executive  summary    Soil   contamination   by   trace   elements   (TE)   is   still   one   of   the   major   environmental  problems   in   Europe.   Consequently,   soil   pollution   has   been   listed   by   the   European  Commission   as   one   of   the   eight   major   threats   to   European   soils.   Soil   remediation   is  therefore   an   urgent   requirement   for   maintaining   one   of   the   most   important   natural  resources   for   maintaining   safe   agriculture   for   producing   food   and   fodder,   but   also  biomass  for  various  other  purposes,  e.g.  production  of  energy  or  raw  materials  such  as  fibre.    Conventional  soil   remediation   technologies  are  very  often  costly,  energy   intensive  and  may   negatively   affect   the   soil   quality   and   functions.   However,   gentle   soil   remediation  options  (GRO)  using  plants,  associated  microbes  and  soil  amendments  may  serve  as  an  environmentally   friendly   and   cost-­‐efficient   alternative.   Although   the   effectiveness   of  GROs  has  been  demonstrated  in  the  lab  and  greenhouse  scale,  information  on  long-­‐term  efficiency   in   the   field   was   still   missing.   Furthermore,   valorisation   options   for   GRO  biomass   and  methods   to   assess   GRO   efficiency  were   hardly   available   and   GROs  were  hardly  considered  in  decision  support  tools  and  remediation  guidelines.    The   main   objectives   of   the   GREENLAND   project   were   to   overcome   the   above   listed  reasons  hindering   the  wider   application  of  GROs.  A  network  of   16   case   studies   on  13  long-­‐term  field  experiments  in  Europe  has  been  established  to  evaluate  and  compare  the  long-­‐term   effectiveness   of   GROs,   including   TE   removal   by   phytoextraction,   or   TE  immobilisation   by   (aided)   phytostabilisation   and   in   situ   stabilization/phytoexclusion.  The  promising  results  were  summarized  in  success  stories  and  published  together  with  a  detailed  guideline   for  GRO   implementation.  These   instructions  also   include   tools   for  enhancing   the   GRO   efficiency   by   choosing   more   efficient   plants,   associated  microorganisms,  mineral  and/or  organic  soil  amendments,  according  to  the  site-­‐specific  conditions.  Since  GROs  often  mainly  affect   the   labile  TE  pool   in  soils,  specific  soil   tests  are   required   for   measuring   GRO   efficiency.   The   GREENLAND   best   practice   guidance  handbook   provides   specific   instructions   for   the   assessment   of   GRO   progress   and  success.    GREENLAND  has  demonstrated  that  GRO  biomass  can  be  valorised  in  several  ways,  e.g.  by   combustion,   anaerobic   digestion   and   pyrolysis,   thereby   providing   energy,   but   also  raw  material  for  further  use,  e.g.  polymetallic  catalysts  for  chemo-­‐catalytic  processes.  In   order   to   enable   stakeholders   and   decision   makers   deciding   for   GROs,   a   decision  support   tool  (DST)  has  been  developed  by  GREENLAND.  This  DST  contains  on  the  one  hand  a  technical  part  that  allows  selecting  the  most  suitable  GRO  for  a  specific  site,  but  on   the   other   hand   it   also   provides   specific   instructions   for   stakeholder   engagement,  which   is   a   key   requirement   for   the   optimal   application   of   sustainable   remediation  strategies  and  in  site  regeneration  more  widely.    The   DST   and   the   best   practice   guidance   handbook   are   available   on   www.greenland-­‐project.eu.    

   2.  Description  of  project  context  and  objectives      Rapid   urbanization   and   industrial   development   has   caused   serious   environmental  problems,  leading  to  contamination  of  soils  in  many  areas  across  Europe.  Soil  is  a  non-­‐renewable  resource,  performing  vital  functions  in  the  biosphere  and  acting  as  a  basis  for  the  majority  of  human  activities   (biomass  and   food  production;   storage  and  cycling  of  nutrients;   water   regulation;   C   sequestration   and   storage;   etc.).   The   European  Environment  Agency   (EEA)  estimates   that  remediation   is   required   for  approximately  250,000   sites   across   Europe   and   that   more   than   80,000   sites   have   been   cleaned   up  during   the   last   30   years   in   EU   countries   for  which   data   on   remediation   are   available.  Among   the   most   common   harmful   contaminants   are   toxic   trace   elements   (also  commonly  called  heavy  metals),  present  in  excess  at  37%  of  the  contaminated  site.    Soil   contamination   due   to   trace   elements   (TE)   in   excess   has   been   identified   by   the  European  Commission  (EC)  as  one  of  the  eight  major   threats   to  European   soils.  Soil  contamination  alters  soil  quality  and  functions,  and  can  negatively  affect  water  quality,  biodiversity,   food   security   and   human   health.   As   part   of   an   initiative   to   address   this  problem  the  EC  adopted  a  Soil  Thematic  Strategy  (COM(2006)  231)  and  proposed  a  Soil  Framework   Directive   (COM(2006)   232)   with   the   objective   of   developing   sustainable  technologies   for  soil  protection  and  remediation  across  the  EU.  However  this  proposal  has   not   yet   reached   full   political   agreement   among   the   Member   States.   Mitigation   of  pollution  and   remediation  of   soils  will  be   crucial   if  European   society   is   to  ensure   that  soil   remains  one  of   its   key  natural   resources.  Remediation   of   soils   has   already  been  applied  for  soil  clean  up,  but  there  have  been  wide  variations  between  Member  States.  It  has   been   estimated   that   the   turn-­‐over   of   the   soil   remediation   industry   in   EU-­‐27  amounted   to   €5.2   billion,   of   which   21.6%   was   spent   in   Germany,   20.5%   in   the  Netherlands,   and  5.9%  each   in  France  and   the  United  Kingdom  (COM(2012)  46   final).  Estimated   clean-­‐up   costs   for   contaminated   soils   across   Europe   are   between   €59   and  €109   billion   (COM(2012)   179   final).   Paganos   et   al.   (2013;  http://dx.doi.org/10.1155/2013/158764)   estimated   that   the   management   costs   for  contaminated  sites  in  the  EU  are  about  6  billion  €  annually.    Over   the   years   several   technologies   have   been  developed   for   reducing   risks   posed  by  polluted  soils.  The  conventional  option  for  remediating  TE  contaminated  soils  has  been  through   “hard”   engineering   approaches   such   as   excavation   and   removal   (often   called  “dig  and  dump”).  However,  many  of  these  techniques  are  cost  and  energy  intensive  and  often   destroy   soil   structure   and   functions.   As   an   alternative,   gentle   remediation  options   (GRO)   have   been   developed   and   proposed   as   a   cheap   and   environmentally  friendly   alternative.   GROs   include   technologies   based   on   the   use   of   plants,   associated  microbes,  and  soil  amendments  for  removing  or  immobilizing  TEs  in  contaminated  soils.  Although  the  GRO  efficiency  has  often  been  demonstrated  in  bench  scale  or  greenhouse  experiments,   results   from   long-­‐term   field   experiments   were   still   limited.   In   order   to  demonstrate   the   GRO   efficiency   in   real   world   conditions   and   to   gain   information   on  associated   issues   (e.g.   biomass   use,   GRO   assessment   and   improvement),   the  GREENLAND  project  has  been  initiated  in  January  2011.      GREENLAND  has  established,  integrated  and/or  extended  16  case  studies  on  13  long-­‐term   field   trials   across   Europe,   covering   different   environmental   conditions   (soil,  

climate,  etc.)  and  different  pollution  scenarios.  Based  on  this  network  of  case  studies,  the    main  objectives  of  GREENLAND  were  as  follows:    

• GRO   option   appraisal   in   the   field   –   comparison   of   GRO   efficiency   and  progress   in   large-­‐scale   and   long-­‐term   field   experiments   under   different  conditions  

o Assessing   the   performance   of   most   promising   GRO   at   large   field   scale  using  the  network  of  the  GREENLAND  case  studies  

o Designing,  preparing,  improving  and  applying  new  and  or  optimised  GROs  o Optimising   the   biomass   production   for   various   valorisation   options   and  

related  ecological  and  financial  returns      

• Assessment  of  valorisation  options  for  biomass  harvested  on  GRO  sites  o Reviewing   existing   processes   and   types   of   biomass   used   in   these  

processes  o Compiling   information  on   current   and  under  development  processes   for  

biomass  cultivated  on  TE  contaminated  lands  o Testing  the  feasibility  of  using  the  various  types  of  plant  biomass  collected  

from  the  case  study  sites,   in  existing  or  under  development  processes   to  assess  advantages  and  limitations  regarding  technical  aspects,  regulations,  acceptance  and  costs.    

• Development  of  a  tool  set  for  monitoring  the  progress  and  success  of  GRO  o Selecting/harmonising  methods   assessing   the   bioavailable/bioaccessible  

TE  fractions  among  the  GREENLAND  case  studies  o Selecting   methods   to   be   used   as   indicators   for   GRO   success   and   as  

sustainability  monitoring  tools.    

• Improving   GRO   through   selection   of   plants   and   associated   microbes   and  modification  in  soil  trace  element  bioavailability  

o Selecting   plant   species,   varieties,   cultivars   or   clones   for   highest   TE  resistance,  TE  extraction  potential,  and/or  biomass  production.  

o Improving  plant  performance  and/or  TE  accumulation-­‐exclusion  potential  using   (a)   conventional   agronomic   practices   or   (b)   biotechnological  approaches  (microbial  inoculants).  

o Identifying   effective   soil   amendments   and/or   amendment-­‐microbial  inoculant  combinations,  for  TE  immobilisation  (phytostabilisation).    

• Development  of  implementation  guidance  and  decision  support  o Preparation  of  a   (multi-­‐lingual)  best-­‐practice  guidance  document   for   the  

application   of   GRO   at   field-­‐scale   (including   appraisal   of   the   various  options   available,   evaluation   of   large   scale   field   trials,   analysis   of  valorisation  potential,  and  suggested  methods  and  monitoring)  

o Development  of  guidelines  for  stakeholder  participation,  engagement  and  empowerment  when  implementing  GRO  

o Developing   and   evaluating   a   decision   support   tool,   focussed   on   GRO,  which   can   be   integrated   into   existing,   well-­‐established   and   utilised  (national)  DSTs  /  decision-­‐frameworks    

 3.    description  of  the  main  S&T  results/foregrounds  (not  exceeding  25  pages),    3.1  Sustainable  management  adapted   to   trace  element-­‐contaminated  sites  and  deployment  of  GRO  at  field  scale    

The   GREENLAND   network   of   field   sites   is   a   cross-­‐European   network   of   metal(loid)-­‐contaminated   sites   where   the   efficiency   of   phytomanagement   strategies   has   been  investigated   on   a   medium-­‐   to   long-­‐term,   under   various   contaminant   (trace   element)  types   and   loadings   and   soil   and   climatic   conditions,   with   various   plant   species   and  cultivars.   The  main   S&T   results   /   foregrounds   delivered   by  WP1   of   the   GREENLAND  project   are   summarized   in   relation   to   WP   deliverables   and   are   more   detailed   in   the  appendix   of   this   document,   as   well   as   success   stories   and   datasheets   of   field   trials  available  on  the  GREENLAND  website  (www.greenland-­‐project.eu).    3.1.1.  Remediation  option  appraisal    The   general   scheme   for   the   remediation   and   phytomanagement   of   trace-­‐element  contaminated   soils   (TECS)   comprises   four   stages:   (1)   risk   assessment,   (2)   option  appraisal,   (3)   implementation   of   remediation   strategy   and   (4)   phytomanagement  (including  biomonitoring  and  maintenance).  Nine  partners  were  deploying  GRO  at  field  scale.   Five   main   types   of   historically   contaminated   sites   were   investigated,   under  different  climatic  and  soil  conditions,  with  either  diffuse  contamination  on  a  large  area  (generally   with   agricultural   soils)   or   local   contaminations   at   mining   sites,   industrial  facilities  and   landfills   (with   technosols  at  several  sites)   (Tab.  3.1.1).  Case  studies  were  categorized  using  a  bioassay  battery  (with  WP3).  Five  exposure  patterns  were  defined  based  on  main  contaminants  involved  in  pollutant  linkages:  Cu  (Touro,  Biogeco),  Cd/Zn  (Pb)   (Lommel,   Bettwiesen,   Phytosed,   Phytoagglo,   Arnoldstein,   Piedrafita,   Högbytorp,  Freiberg),  As  (Reppel,  Jales,  Freiberg),  Cr/Mo  (Rive  de  Gier),  and  metal/PAHs  (Biogeco,  Chaban-­‐Delmas,  Phytoagglo,  Borifer).    Tab. 3.1.1 Summary of the Greenland network investigating GRO in long-term field trials Partner Sites Country Sources GRO Main contaminants

Landfill PT-F Bettwiesen Switzerland former hot dip 1 Cd, Zn, Pb Zn galvanizing plant IUNG Piekary Poland Cd/Zn/Pb tailings 3 Zn, Cd, Pb INRA Chateauneuf France steel mill wastes 3 Cr, Mo, Zn, Ni, Cu, Pb

Atmospheric depositions on a large agricultural area HAU Lommel Belgium Zn/Pb smelter 1 Cd, Zn, Pb AIT Arnoldstein Austria Zn/Pb smelter 3 As, Cd, Zn, Pb LfULG Freiberg-Halsbrücke Germany Zn/Pb smelter 1, 2 As, Cd, Pb HAU/INRA Reppel Belgium/France As refinery 1 As, Zn, Pb, Cd

Wood preservation facility INRA Biogeco France wood preservative 1, 2, 4 Cu, Cu/PAHs

Mine, tailings CSIC Touro Spain Cu mining 1, 2 Cu CSIC Piedrafita, Rubiais Spain Pb/Zn mining 1 Zn, Pb, Cd INRA Jales Portugal/France Au mining 2, 3 As, Zn

Technosols, other sources SLU Högbytorp French trial Sweden Irrigation with landfill 1 (Cd), Cr, Zn leachates

INERIS Phytosed ech 1 France dredged sediments 1, 2 Zn, Cd, Cu, Cr, Pb, Mo INERIS Phytagglo France brownfield 1, 4 Zn, Cd, PAHs INRA Chaban Delmas France embankments, harbor 2, 4 Zn, Pb, Cu, Cd, Ni, As, Hg, PAHs facilities INRA Borifer France metal surfacing 2, 4 Zn, Pb, Cu, Ni, As, Hg, PAHs ________________________________________________________________________________________________________________________________ 1: (aided) phytoextraction, 2: (aided) phytostabilization, 3: in situ immobilization/phytoexclusion, 4: rhizodegradation  3.1.1.1.  Setting  of  conceptual  models    

Conceptual  models  (CM)  were  built  for  all  sites  (n=16)  summarizing  main  sources  of  soil  contamination,   exposure   to   TE   and   soil   ecotoxicity,   biological   receptors   such   as   plant  and  microbial   communities,   initial   pollutant   linkages   and   risks   on   site   and   nearby.   At  five  sites,  vertical  migration  from  the  topsoil  as  well  as  wind  erosion  were  considered.  Clusters   with   different   soil   ecotoxicity   were   defined   at   most   sites.   Diverse   physico-­‐chemical  parameters,  soil  and  technosol  types  were  of  concern,  from  acid  sandy  soils  to  calcareous   soils   and   alkaline   technosol.   Mixed   contamination   (TE   and   organic  contaminants  such  as  PAHs)  was  taken  into  account  at  4  sites.  Climatic  conditions  varied  from   cold   climate   and   short   summer   (Sweden)   to   Mediterranean   climate   and   dry  summer  (Spain).    Different   end   land   uses   were   considered:   landscaping,   recreation   area,   production   of  annual   crops   for   (non-­‐food)   plant-­‐based   feedstock   and   biosourced   chemistry,  production   of  metal-­‐excluder   crops   (grasses,   cereals),   production   of  wood   from   short  rotation  coppice  (SRC)  and  tree  planting  in  line  with  eco-­‐technologies.  .      3.1.1.2.  Risk  assessment  –  pollutant  linkages    

Spatial  distribution  of  soil  contaminants,  soil  physico-­‐chemical  parameters,  labile  soil  TE  and  ecotoxicity  were  characterized  (with  WP3).  Initial  and  residual  risks  and  pollutant  linkages   were   quantified,   in   line   with   the   GRO   implementation   and   their  (bio)monitoring.  The  TE  concentrations  in  harvested  plant  biomasses  were  determined  at   each   harvest   for   all   sites,   and  were   compared  with   the   legislation,   common  values,  and   the   needs   of   local   conversion   chains.   Concentrations   of   TE   in   leachates   from   the  topsoil,  changes  in  labile  TE  pools  in  the  soils  and  residual  risks  for  plants  and  microbial  communities   were   generally   monitored   over   four   years.   Organization   of   plant  communities   and   interspecific   variability   of   TE   concentrations   in   plant   parts   were  investigated.    3.1.1.3.  Option  appraisal    

This  stage  aims  at  establishing  which  remediation  option,  or  combination  of  options,  can  alleviate  all  pollutant  linkages  that  present  an  unacceptable  risk  at  the  site.  It  includes:  identification   of   feasible   remediation   options,   detailed   evaluation   of   options,   and  developing   the   remediation   strategy.   The   Greenland   project   added   another   aim:   the  implemented  GRO  must   improve  the  ecosystem  services,  notably  provisioning  services  through  biomass  production  for  the  bio-­‐economy  and  other  ecosystem  services  such  as  carbon   sequestration,   recycling   of   organic   matters,   water   filtration,   quenching   of   soil  erosion  and  restoration  of  plant-­‐microbe  communities,  without  generating  wastes  and  pollutant   linkages.  The  main  concerns  were   to  ensure   that   remediation  option  criteria  selected   for   the   soil   are   protective   for   controlled   waters,   plant,   microbe   and   animal  communities,   and   in   compliance   with   current   legislation   for   labile   (extractable)   TE  fraction   in   the   soil,   forages   and   feedstuffs,   foodstuff,   groundwater  or   on  upper   critical  threshold   values   according   to   experts.   To   better   determine   the   benefits   and   limits   of  feasible  GRO   for   some  or   all   clusters   at   one   site,   according   to   the   selected   conceptual  

scheme  and  end   land  use,   it   is   recommended   to   compare   them  with   the  best   relevant  conventional  remediation  options,  in  parallel  in  pot  and  field  experiments  (with  similar  soil   contamination).   In  case  of   failures  of  GRO   in   the   long-­‐term,   the  other   remediation  options  would   be   deployed   on   the   site   clusters.   Soil   amendments   and   plant  materials  were  investigated  in  coordination  with  WP4.    Option  appraisals  resulted  in  the  selection  of:  ●  phytoextraction  at  13  field  trials  (5  with  SRC,  5  with  high  yielding  crops,  and  3  with  hyperaccumulators),    ●   (aided)   phytostabilisation:   11   field   trials   (3   with   SRC   and   8   with   perennial  herbaceous  plants)    ●  in  situ  immobilization/phytoexclusion:  5  field  trials.      3.1.2.  Implementation  of  remediation  strategies    Most  GRO  were   selected  based  on  pot   and  mesocosm  experiments   (option   appraisal).  Main  lessons  gained  on  GRO  implementation  were:    

●   quantify   the   spatial   variability   of   parameters   driving   the   choice   of   feasible   GRO  according   to   the   current/future   land   uses   (for   each   cluster)   and   the   related  target/trigger   values   (notably   those   from   the   legislation   and   exposome)   and   other  drivers  (land  value,  time  constraints,  etc.).    ●  account  for  any  specific  requirements  related  to  the  selected  feasible  GRO  and  the  best  conventional  option  (to  be  compared).  ●   compare   the   best   conventional   technology(ies)   in   parallel   with   the   selected   GRO  emerging  from  option  appraisal.    ●  don’t  upscale  directly  from  ‘pot  experiments’  to  ‘full-­‐scale’  (in  situ)  deployment  on  the  cluster(s)  without  the  return  skill  of  biomonitoring  and  maintenance  for  several  years.    ●   set   fences   around   small   clusters,   especially   at   the   start   of   the  phytomanagement,   to  protect  the  trees  and  other  attractive  plant  species;  ●   define   reasonable  plot   size   for  avoiding   the  edge  effects  and  permitting  a   long-­‐term  (>5  years)  monitoring  and  maintenance.,    ●  don’t  forget  to  monitor  the  foliar  exposure  ●  adopt  appropriate  agronomic  practices:    ●  the  choice  of  initial  plant/microbe  partnerships  must  account  for  the  local  conversion  chains   of   biomass   (generally   the   biomass   production   on   one   rather   small   site   is   not  enough  to  financially  support  a  dedicated  local  valorization  plan;  this  biomass  must  be  commonly  merged  with  similar  biomass  from  other  sites  (forest,  SRC,  agricultural  field,  green  wastes,   etc),   provided   that   their   composition   is   suitable  with   the   process   or   its  marketing   image.   Plants  must   not   only   show   tolerance   to   the   contaminant(s)   present  but  also  resist  other  abiotic  and  biotic  potential   stresses,  e.g.  water  stress,   soil  acidity,  frost,  soil  erosion/compaction,  herbivory,  pests,  nutrient  deficiency,  salinity,  etc.    Try   to   implement   the   young   trees   before   to   implement   the   herbaceous   crops  underneath  and  in  between.  It  is  pivotal  to  irrigate  trees  in  year  1  (and  sometime  year  2)  during  dry  periods  to  increase  the  survival  rate  and  promote  the  establishment  of  their  root   systems   (depending   on   soil   type,   climatic   conditions,   etc.).   Pay   attention   to   the  slope,   potential   soil   erosion   and/or   flooding.   In   case   of   excluder-­‐based   SRC   for  bioenergy  purposes,  the  selection  of  genotypes  can  be  based  on  their  characteristics  in  line  to  conversion  processes,  e.g.  calorific  value,  bulk  density,  moisture  content,  ash  and  extractive  content.  Transplantation  of  mycorrhizal  trees  was  more  successful  than  that  of  non-­‐mycorrhizal  ones  and  the  on  site  mycorrhization  of  tree  cuttings.    

Salix  and  Populus  clones  show  high  variations  in  biomass  production,  TE  tolerance  and  accumulation  patterns  in  roots,  leaves,  and  even  in  wood  between  clones.  Some  species  and  clones  of  willow  have  high  bioconcentration  factors  (BCFs)  for  Cd  (up  to  27)  and  Zn  (up  to  3).  Given  the  ample  variation  in  metal  accumulation,  best-­‐performing  clones  can  be   selected   based   on   their   TE-­‐tolerance,   uptake   efficiency   (accumulating   clones   for  phytoextraction  vs.  excluding  clones  for  phytostabilisation),  translocation  from  roots  to  shoots,   and  biomass  production.  Clones  can  be  selected   for   their  ability   to  accumulate  certain  metals  (e.g.  Cd  and  Zn)  while  at  the  same  time  immobilizing  elements  such  as  Cu  and   Pb.   Evidences   of   tolerance   to   TE   and   fungal   and   insect   infection,   e.g.   leaf   rust  (Melampsora  sp.)  and  lace  bug  (Monosteira  unicostata  Muls.  and  Rey),  cold  and  drought  adaptation  were  revealed  at  the  Lommel  site.    ●  combine  phytomanagement  and  ecology:  establish  natural  and  passive  habitats  to  host  and   promote   reproduction   of   the   biological   auxiliaries   (notably   beneficial   insects   and  birds)   and   counteract   bioagressors.   Think   about   the   connection   of   clusters   with   the  other   ones   nearby.   Use   corridors   allowing   the   predators   (fox,   raptors,   etc.)   to   hunt;  these  corridors  can  be  combined  with  the  access  required  for  monitoring  and  sampling  as  well  as  the  harvest  machines.  Avoid  a  full  site  monocultures  to  alleviate  the  selection  of  pest  populations   (e.g.  use  diverse  clones/genotypes   for   trees   in  clusters;  use  a  crop  rotation  in  case  of  annual  plants.  ●   Phytomanagement   can   combine   some   GRO:   The   phenotype   of   plant   species   in  response  to  TE  excess  is  element  dependent  and  a  plant  assemblage  can  support  various  GRO   at   the   same   time   on   mixed-­‐contaminated   soils.   For   example   a   poplar   SRC   can  simultaneously  phytostabilize  Cu/Pb  in  its  root  system,  phytoextract  Cd/Zn  in  its  aerial  parts  and  promote  the  rhizodegradation  of  xenobiotic  organic  compounds.      

3.1.3.  GRO  implemented  and  biomass  production    3.1.3.1.  In  situ  immobilization/phytoexclusion    This  GRO  can  be   implemented  as  either  a   long-­‐lasting  (phyto)management  option  or  a  temporary,   reversible   one   that   can   be   later  modified   based   on   the  monitoring   results  from  the  phytomanaged  plots.  Decreasing  the  labile  TE  pools  in  TECS  by  incorporation  of  soil  conditioners  and  the  use  of  excluder  plants  are  both  main  approaches.  Different  soil   conditioners  were   investigated  on  a   long-­‐term  and  at   field   scale,   i.e.   phosphates,  composts  and  technosols,   iron  bearing  materials  (iron  grit,  gravel  sludge),  and  alkaline  materials   such   as   alumino-­‐silicate   slags,  marl   lime,   biosolids,   and   dolomitic   limestone  (see  Appendix)    End  land  use:  annual  crop  production  Staple  crops  and  oilseeds:  Selection  of  efficient  excluder  cultivars  of  wheat,  barley,  rice,  potato   and   maize   for   cultivation   on   contaminated   and   remediated   land   contributes  towards   reducing   the   entrance   of   non-­‐essential   TE,   and   also   avoiding   the   excess   of  essential  ones,  into  the  food  chain.  Cd  is  of  highest  concern  regarding  metal  uptake  into  the   food   chain   as  well   as   As,   Mo,   Se,   Tl   and   Hg.   Excluder  maize,   barley   and   potatoes  cultivars  were  long-­‐term  assessed  at  Arnoldstein:  use  of  the  excluder-­‐phenotype  Bodega  vs.   accumulator-­‐phenotype   Hellana   reduced   barley   grain   Cd   by   over   40   %.   In  combination  with  the  incorporation  of  gravel  sludge  and  red  mud  into  the  contaminated  soil,  a  further  >30  %  Cd  uptake  could  be  avoided.      Crop  rotation  included  winter  oilseed  rape,  winter  wheat  and  spring  barley  at  Freiberg-­‐Hilbersdorf  in  combination  with  marl  lime  application.  Soil  pH  was  slightly  changed  by  lime   and   P   application   and   was   generally   increased   in   year   4,   which   resulted   in   a  

decrease  of  mobile  Cd  (by  50-­‐75%).  Based  on  BCF  of  Cd,  Zn  and  Pb,  the  barley  cultivar  “Salome”   was   shown   to   less   accumulate   these   metals   in   its   grains   than   the   Marthe  cultivar.  Based  on  EU  directive  2002/32/EC  (2002)  these  barley  grains  were  suitable  as  single   fodder.   Considering   changes   in   element   transfer   into   plant   parts   as   affected   by  amendment  options,  grain  Pb  differed  between  the  control  and  P  treatment  with  highest  concentration   and   the   limed   treatments   with   low   concentration,   especially   for   the  combined  fertilized  treatment.  The  biomass  production  of  winter  oilseed  rape  for  both  cultivars  was  within  the  common  range  of  yields  for  this  German  region  (2.4  –  4.4  t/ha).  Those   of   winter   wheat   were   below   the   range   (5   –   8   t/ha),   especially   for   the   low  accumulating   cultivar   Türkis,   which   produced   a   lower   grain   yield   than   to   the   high  accumulating  cultivar  Tiger.  The  grain  yields  of  spring  barley  were  below  the  common  range  (4.2  –  7.4  t/ha)  with  slight  differences  between  both  cultivars.      Grassland  management:   Grassland   based   on   TE   excluder   grassy   crops   is   one   relevant  GRO   to   alleviate   windblown   dust   and   water   runoff   on   large   TE-­‐contaminated   areas,  notably  with  low  fertility.    At   Arnoldstein,   shoot  DW  yield   reached   5   t/ha/yr.   The  most   efficient   soil   conditioner  (gravel  sludge  and  red  mud,  slurry  management)  was  reducing  the  labile  pools  of  Zn  (-­‐90%),   Cd   (-­‐80%),   and   Pb   (-­‐90%)   in   the   soil.   Plant   monitoring   based   on   Plantago  lanceolata   indicated  reduced  shoot  concentrations  for  Cd  (-­‐70%)  and  Zn  (-­‐77%).  Shoot  Cd   and   Pb   concentrations   of   harvested   grass   mixture   just   exceeded   the   maximum  permitted  concentrations  (MPC)  in  forages.      At  Freiberg-­‐Hilbersdorf,  soil  pH  (CaCl2)  varied  between  4.3  and  5  in  the  unamended  soil.  It   reached  pH  6   at   2t  marl   lime/ha   and  6.5   at   4t/ha.   Extractable   soil   Cd  was   reduced  from  0.6-­‐0.7   to  0.05-­‐0.1  mg/kg  soil  DW.  Grass   shoot  DW  yield  varied   from  2   to  5.75   t  DW/ha  depending  on  season  (3  cuts/yr)  and  soil  amendments.  It  was  enhanced  in  May  by   marl   lime   application   at   4t/ha.   Shoot   As   concentrations   of   grass   did   not   differ  between  unamended  and  2t/ha-­‐amended  soils  on  the  2012-­‐2013  period  and  was  in  the  0.25-­‐0.5   mg/kg   DW   range.   It   started   to   decrease   on   the   third   year.   Shoot   As  concentration  was  higher   in   the  4t/ha-­‐treated  soil,   reaching  1.5-­‐1.75  mg  As/kg  DW   in  2012-­‐2013,   despite   high   shoot   DW   yield,   and   also   decreased   to   0.6  mg   As/kg   DW   in  2014  after  the  last  marl  lime  application.  Shoot  Cd  concentration  ranged  between  1-­‐1.5  mg/kg   DW   in   2012   and   did   not   differ   across   the   treatments.   It   decreased   in   all  treatments   in   2014   (0.5-­‐0.8   mg/kg   DW),   but   lower   values   in   marl   lime-­‐treated   soils  were   similar   to   the   unamended   soil.   Shoot   Pb   concentration   varied   from   0.3   to   4  mg  Pb/kg  DW  in  average  and  reached  8  mg/kg  DW  in  some  shoot  samples  from  the  4t/ha-­‐amended  plots.  It  was  decreased  in  year  3  for  all  plots  with  a  lower  value  at  2  t/ha  (0.3  mg  Pb/kg  DW)  compared  to  the  unamended  soil.      

At  Piekary  (PL):  The  grass  mixture  consisted  of  local  cultivars:  Festuca  rubra  L.  cv.  Atra,  Poa  pratensis  L.  cv.  Alicja,  Festuca  arundinacea  Schreb.  cv.  SZD,  and  Festuca  ovina  L.  cv.  Sima.   17   years   after   biosolid   incorporation,   water-­‐soluble   fractions   of   major  contaminants  (Zn,  Cd,  and  Pb)  in  the  soils  remained  at  low  levels,  in  line  with  soil  pH  and  Ca-­‐carbonate   distribution   over   the   field.   Soil   bacterial   communities   were   highly  diversified   in   amended   soils.   Dehydrogenases   activity   increased   as   water   extractable  metal  (Cd,  Zn)  fractions  in  the  soils  were  reduced.  Plant  cover  and  biomass  production  depended   on   the   soil   treatment   being   highest   soils   amended  with   biosolid   combined  with  byproduct  lime.  Most  persistent  grass  species  were  P.  pratensis,  Agrostis  capillaris  and  F.  ovina.  These  species  covered  the  largest  area  of  the  field.  A  substantial  part  of  the  areas  was  covered  by  colonists:  Calmagrostis  epigejos,  Hypochoeris  radicata,  Melandrium  

album,   Artemisia   vulgaris,   Daucus   carota   and   Solidago   gigantean.   Untreated   tailings  outside  the  reclamation  area  remained  barren.    3.1.3.2.  (aided)  phytostabilisation    3.1.3.2.1.  SRC  with  and  without  grass  cover/herbaceous  layer    

SRC  parameters:  Many  tree  species  are  suited  for  phytostabilization  due  to  their  deep  root  systems,  high  transpiration  rate,  high  TE  tolerance,  and  ability  to  grow  on  nutrient-­‐poor   soils.   Trees   can   stabilize   less   mobile   metals   (e.g.   Cu,   Pb,   and   As)   in   the   soil   by  physically   preventing   migration,   leaching,   and   soil   dispersion;   alternatively,   they   can  immobilize  TE  through  uptake  and  accumulation  by  the  roots  into  the  plant,  adsorption  on  the  root,  and  precipitation  in  the  rhizosphere.    At  Biogeco  (Cu-­‐contaminated  soils)  fertilized  mycorrhizal  poplars  were  harvested  two  times,  whereas  minimum  values  of  potential  biomass  to   initiate  the  harvest  of  willows  and  non-­‐mycorrhizal  poplars  were  not  reached.  For  willow  SRC,  it  can  be  done  in  year  9  only   for  ectomycorrhizal   trees.   Shoot  DW  yield  of  poplar  SRC  varied   from  20   to  270   t  DW/ha  showing  the  spatial  variability  of  soil  exposome,  fertility  and  water  supply,  135  t  DW/ha  was  even  reached  in  some  untreated  plots  nearby  other  fertilized  plots  managed  by  phytoextraction.  In  year  6  after  amendment  incorporation,  compost  (OM)  increased  poplar   growth   compared   to   the   untreated   soil   (UNT),   whereas   addition   of   dolomitic  limestone  (DL)  resulted  in  less  significant  increases.      At  Phytosed,   two  willow  cultivars  (Tordis  and  Inger)  were  planted.  Their  survival  rate  in  year  2  was  89%  accounting  for  all  plots,  but  it  dropped  to  75%  for  Tordis  in  several  amended  plots  and  the  rate  of  chlorotic  leaves  reached  30-­‐50%  in  these  plots.  The  foliar  Cd  concentrations  were  high  (10-­‐30  mg  Cd/kg),  the  Tordis  willow  clone  showing  higher  values   than   the   Inger   in   both   the   Thomas   basic   slag   (TBS)-­‐amended   plots   and   in   the  control  plots.  Similarly  foliar  Zn  concentrations  ranged  from  ~  2000  to  3500  mg  kg-­‐1  DW  whereas  common  values  varied  from  81  to  296  mg  kg-­‐1  DW.  Wood  and  bark  Zn  and  Cd  concentrations  in  year  2  (2  mg  Cd/kg  DW)  were  lower  than  those  in  willow  leaves.  Bark  concentrations   (10-­‐15   mg   Cd/kg)   were   higher   than   wood   concentrations   and  concentrations   increased   with   the   height   of   willow   due   to   the   increase   of   bark  proportion.  Tordis  willows  accumulated  more  Cd  than  Inger,  in  accordance  with  the  leaf  results.   The   alkaline   amendment   did   not   reduce   or   at   least   stabilize   the   TE  concentrations  in  aerial  plant  parts  of  willows.    Herbaceous  layer:  at  Phytosed  (FR)  a  commercial  alkaline  by-­‐product  of  steel  industry  used  in  agriculture  was  incorporated  (rate  9  t  ha-­‐1,  pH  8)  into  the  technosol  for  reducing  the  metal  mobility  and  promoting  the  grassy  crop.    After   2   years,   the   commercial   cultivar  Barchampsia   cespitosa   is   a   good   candidate   for  phytostabilisation  (i.e.  success  of  the  plant  cover,  tolerance  to  the  technosol  conditions,  shoot  TE  concentrations   close   to   common  values   for  grasses  on  uncontaminated   soil).  This   grass   competes   well   against   the   invasive   species   (beneficial   effect   of  phytostabilisation;  see  previous  report).      Responses   to   bioagressors:   Both  willow   cultivars   Inger   and   Tora   at  Phytosed   were  susceptible  to  the   imported  willow  leaf  beetle  (Plagiodera  versicolora).  A  similar  biotic  interaction  was   occurring   at   Lommel.   The  use   of   native   poplar   and  willow   at  Biogeco  reduced  disease  incidence,  particularly  from  Melampsora  rust.    

Touro   (ES):   the   Cu-­‐mine   tailings   were   amended   with   three   mixtures:   composted  municipal  solid  wastes  (compost)  and  two  technosol  mixtures  and  planted  with  metal-­‐tolerant   clones   of   Salix   (S.   caprea   and   S.   viminalis)   and  Populus  nigra,   or  with   a   grass  cover  Agrostis   capillaris   cv.   Highland.  Mortality  was   high   on   technosol-­‐amended   plots  but  low  on  compost-­‐amended  plots.  Growth  and  survival  (70-­‐80%)  of  woody  trees  was  optimal   in   compost-­‐amended   plots.   After   three   years   tree   height   was   highest   in   S.  viminalis  and  P.  nigra  (reaching  up  to  3-­‐4  m).    Changes  in  soil  exposome  /  TE  mobility  in  soils  Touro:  In  year  3,  soil  NaNO3-­‐extractable  Cu  concentrations  remained  low  (<1  mg/kg)  in  all   treated  soils  without   influence  of   the  vegetation  cover   type.  Soil  pH  was  3.5  before  GRO  implementation,  and  in  year  3  remained  between  6.0  and  7.0  in  compost-­‐amended  soils.  Soil  pH  was  higher  in  soils  under  Salix,  followed  by  Agrostis  and  finally,  unplanted  soils.  Phytosed:   In  year  2  extractable  Zn  and  Cd  fractions  (roughly  0.4-­‐0.8  mg  Zn  and  0.001-­‐0.0015  mg  Cd/kg  soil)  did  not  differ  between  the  amended  and  unamended  technosols.  Biogeco:  In  year  5,  Cu  concentration  in  the  soil  pore  water  was  higher  in  the  compost-­‐amended   soils   for   both   mycorrhizal   and   non-­‐mycorrhizal   trees   and   the   lower   in   the  limed  soils  with  mycorrhizal  trees.    Restoration   of   soil   microbial   activity   and   communities:   At   Touro,   soil   enzyme  activities   (involved   in  C,  N  and  P  cycles)  were   lowest   in  untreated  soils  and   increased  with  time  in  amended  soils.  A  plant-­‐induced  effect  was  observed:  activities  were  higher  in  plots  planted  with  woody  trees,   followed  by  Agrostis,  and  lowest  in  unplanted  plots.  Shifts   in  the  structure  of  the  soil  bacterial  community  (total  Eubacterial  community,  a-­‐  and  b-­‐proteobacteria,  Actinobacteria  and  Streptomycetaceae)  were  compared  over  time.  At   each   sampling   period   (after   1,   2   and   3   years)   the   bacterial   community   of   soils  sampled   before   GRO   implementation   differed   from   that   of   phytomanaged   soils.   At   all  sampling  periods,  the  soils  amended  with  compost,  technosol  1  or  technosol  2  differed.  In  general  the  soil  bacterial  communities  continue  to  cluster  separately  according  to  the  plant  species,  either  Agrostis  or  Salix  cultivation.  There  was  a  trend  towards  an  increase  in   bacterial   diversity   with   time,   and   also   a   higher   bacterial   diversity   in   planted   soils  (albeit  Agrostis  or  Salix)  compared  to  unplanted  (but  amended)  soils.      3.1.3.2.2.  Grassy  crops  (only)    

TE   excluder   perennial   herbaceous   crops   such   as   switchgrass   (Panicum   virgatum),  Miscanthus  spp.  and  vetiver  (Vetiveria  zizanioides)  have  wide  climatic  adaptability,   low  production   costs,   suitability   to  marginal   lands,   relatively   low  water   requirements,   low  nutrient   and   agrochemical   needs,   and   potential   environmental   benefits.   They   can  provide   feedstock   for   the   energy   sector   or   essential   oils.  Miscanthus   x   giganteus   and  Vetiver  were   implemented   at  Biogeco.   For  Vetiver   in   year   4,   shoot  DW  yield   reached  38t/ha   in  uncontaminated  plots,  7-­‐15t/ha   in   the  amended  Cu-­‐contaminated  plots,  and  only  2.6   t/ha   in   the  highly  Cu-­‐untreated  plots,  as   total  Cu   in  soil  pore  water   increased  from   0.2   to   0.9  mg   Cu/L.   It   always   demonstrated   a   Cu-­‐excluder   phenotype,   shoot   Cu  concentration   being   in   the   10-­‐13   mg/kg   range   with   no   influence   of   soil   Cu  contamination.  For  Miscanthus,  shoot  DW  yield  in  year  3  varied  from  0.07  (Unt)  to  1.8  (compost  and  dolomitic   limestone,  OMDL)  t  DW/ha.  Its  shoot  Cu  concentration  ranged  from   7   (OMDL)   to   95  mg   Cu/kg   DW   (Unt).   The   single   incorporation   of   compost   and  dolomitic  limestone  still  reduced  labile  soil  Cu  and  Miscanthus  exposure  in  year  7.  Shoot  Cu  removals  in  year  3  varied  from  3  to  17  g  Cu/ha  depending  on  shoot  Cu  concentration  

and  shoot  DW  yield,  with  maximum  at  median  soil  Cu  contamination  and  minimum  for  soil   Cu/PAH   contamination.   For   Sorghum   (Sorghum   spp.)   both   cultivars   for   biomass  and  bioenergy  were  not  successful  at  field  scale,  being  too  sensitive  to  Cu  excess  and  low  water  supply  in  sandy  soils.    Grassland:   At   Biogeco   compost   (OM)   and   dolomitic   limestone   (DL),   singly   and   in  combination  (OMDL)  were  assessed.  In  year  7,  most  plots  initially  planted  with  grasses  were   dominated   by   an   assemblage   of   Cu-­‐tolerant   Agrostis   capillaris   and   A.   gigantea  whatever   the   soil   treatments.   Other   introduced   grassy   species   such   as   Sporobolus  tenacissimus   and   B.   cespitosa   are   disappearing.   Shoot   DW   yields   were   higher   in   the  compost-­‐amended   plots   compared   to   the   limed   ones.   Shoot   Cu   concentration   was  slightly   lower   for   the   grass   species   harvested   in   the   OMDL   plots.   Highest   shoot   DW  yields   in   the   compost-­‐amended   plots   led   to   maximum   shoot   Cu   removals.   Total   Cu  concentration   in   the   soil   pore   water   was   increased   in   the   compost-­‐amended   soils   as  compared  to  the  limed  soils.    Chaban-­‐Delmas   (FR):   A   total   of   72   plant   species   were   identified   in   the   grassland,   32  species  being  occasionally  present.  Main  species  discriminated  three  plant  subsets,  one  dominated   by   Medicago   sativa   and   Lolium   perenne   with   the   lowest   plant   species  richness,   one   dominated   by  Arrhenatherum  elatius,  Bromus  sterilis,  Holcus   lanatus   and  Dactylis  glomerata,  with  a  median  value  for  the  species  richness,  and  one  Melilotus  albus,  Trifolium   arvense,   and   Trifolium   pratensis   as   dominant   with   the   highest   bare   soil  percentage  but  the  highest  species  richness.    Touro   (ES):   Agrostis   capillaris   cv.   Highland   was   successfully   established   in   both  compost-­‐   and   technosol-­‐amended   plots.   Shoot   Cu   concentrations  were  within   normal  levels  for  grass  species  growing  in  uncontaminated  soils,  and  lower  than  in  grass  species  colonizing   the   surrounding  untreated   tailings.  Nutrient   concentrations  were   increased  in  all  amended  soils,  notably  in  technosol-­‐amended  plots.  Shoot  DW  yield  was  highest  in  compost-­‐amended  plots.    3.1.3.2.3.  (Aided)  phytoextraction      The  aims  were  to  quantify  (1)  the  biomass  production,  (2)  the  plant  ionome  (notably  TE  concentrations),   and   (3)   the  TE  phytoextraction.  Additional   aims  were   (4)   to   improve  agronomic   practices,   (5)   to   enhance   ecosystem   services   such   as   C   sequestration   and  microbial  activities,  and  (6)  to  create  economic  opportunities  from  the  biomass.  Plants  must  be  able  to  accumulate  high  TE  concentrations  in  their  harvested  parts  and  have  a  reasonably  high  biomass  production.  Relevant  options  were  TE-­‐hyperaccumulators  and  secondary   TE   accumulators.   Three  main   options   were   implemented   to   address   three  mains   situations,   Cd/Zn   (Pb),   Cu   (Cu/PAH)   and   As/metal   excess:   high-­‐yielding   crops  (HYC),   short   rotation   coppice   (SRC),   and   herbaceous   hyperaccumulators,   in  monoculture   and   co-­‐cropping.   Influences   of   soil   conditioners   such   as   compost,   Linz-­‐Donawitz  slags,  soil  acidifying  agents  (citric  acid,  S),  and  co-­‐cropping  were  investigated  to  enhance  TE  phytoextraction.      3.1.3.2.3.1.  High-­‐yielding  crops  (HYC)    

High-­‐yielding  crops  (annuals  or  perennials)  are  recognized  as  viable  alternatives  for  TE  phytoextraction   (particularly   Cd,   Se   and   Zn)   if   they   show   relevant   shoot   TE   removals  (i.e.   moderate-­‐high   BCF   and   high   shoot   yield).   In   vitro   breeding   and   chemical  mutagenesis   can   improve   the   metal   tolerance   and   phytoextraction   capacity   of   high-­‐

yielding  crops  such  as  tobacco  and  sunflower  (see  WP4).  At  the  Rafz  site  (Switzerland),  shoot  metal  removals  by  the  sunflower  mutants  were  up  to  7.5-­‐,  9.2-­‐  and  8.2-­‐fold  higher  for   Cd,   Zn   and   Pb   than   the   inbred   line,   respectively.   As   monocultures   can   lead   to   a  decline  in  biomass  yield  due  to  the  depletion  of  nutrients,  occurrence  of  diseases,  pests,  and   weeds,   and   have   a   negative   effect   on   soil   fertility,   crop   rotations   such   as  sunflower/tobacco  (with  winter  fodder  pea  at  Bettwiesen,  and  white  clover  at  Biogeco  as   cover   crops   during   winter   for   green   manure   and   limiting   soil   erosion)   were  investigated.   Fibre   hemp   (Cannabis   sativa)   and   kenaf   (Hibiscus   cannabinus)   were  cultivated   at   Lommel.   Tobacco   and   sunflower   mother-­‐clones   and   variants   were  cultivated   at   5   sites.   Datasets   are   available   for   shoot   DW   yields,   shoot   metal  concentrations   and   shoot  metal   removals.   The   influences   of   soil   contamination   levels,  fertilization,   maintenance   through   compost   dressing,   plant   species   and   genotypes  (mother-­‐line,   somaclonal   tobacco   variants,   and   sunflower   mutants),   agronomic  practices  such  as  irrigation,  co-­‐cropping  and  flower  topping  were  considered.    Touro   (ES):  Shoot  DW  yield  of  tobacco  primarily  depended  on  the  climatic  conditions  during   the   growth   season:   the   highest   biomass   for   all   genotypes  was   achieved   in   the  2014  harvest  and  reached  3400-­‐4000  kg  DW  ha-­‐1  (levels  comparable  to  those  obtained  in   the   agricultural   soils   of   Bettwiesen   and   Lommel).   Differences   between   the   BAG  motherline  and  the  10-­‐8  and  10-­‐4  variants  were  not  pronounced,  and  biomass  tended  to  be   higher   for   BAG.   Shoot   Cu   removal   in   2014   (60-­‐70   g   Cu/ha)   was   lower   than   that  obtained  in  Biogeco.    Piedrafita   (ES):   Sunflower   could   only   grow   in   the   compost-­‐amended   plots,   while  tobacco   could   grow   in   both   compost-­‐amended   and   untreated   mine-­‐soils   (after  fertilization   with   inorganic   NPK).   Tobacco:   Annual   shoot   DW   yield   varied   widely  according   to   climatic   conditions   and   to   competition   with   weeds;   biomass   production  was   highest   in   2012.   There   were   no   consistent   differences   between  motherlines   and  other  variants.  Biomass  production  and  Cd/Zn  extraction  potential  was  lower  than  that  observed   in   Lommel   and   Bettwiesen.   Sunflower:   its   cultivation   was  more   successful  than   tobacco  at   this   site.  Annual   shoot  DW  yields  were   similar   to   that  obtained  at   the  other  Greenland  field  sites.  Cd/Zn  extraction  potentials  were  similar  to  that  obtained  in  Bettwiesen   and  Lommel   in   the   2012   and  2013  harvests,   but   lower   in   2014.  Mutant   1  reached  up  to  6772  g  Zn  and  23  g  Cd  /ha  extraction  potential.    Biogeco  (FR):    Tobacco:  its  shoot  DW  yield  depended  on  total  soil  Cu,  soil  amendments,  the  genotype  in  some  plots,  and  annual  climatic  conditions.  Flower  topping  in  years  6  and  7  allowed  the  development   of   bottom   suckers,   which   increased   the   shoot   biomass.   Depending   on  climatic   conditions,   early   flower   topping   in   Southwest   France   allowed   to   harvest  tobacco  shoots  two  to  three  times  per  year  and  to  avoid  loss  of  dried  leaves.  ●  At  moderate  soil  Cu  contamination  (258  –  382  mg  Cu/kg):  in  year  6,  shoot  Cu  removal  by  the   OMDL   plants   reached   84-­‐132   g   Cu/ha,   without   genotype   influence.   The   second  compost  dressing  in  year  6  (OM2DL)  reduced  shoot  Cu  removal  as  compared  to  OMDL,  likely  due  to  a  decrease  in  mobile  soil  Cu.  However,   in  year  7,  the  OM2DL  plants  had  a  higher   shoot   yield   than   the   OMDL   ones,   likely   due   to   nutrients   released   by   compost  mineralization.  ●  At  high   soil  Cu   contamination   (894  –  1020  mg  Cu/kg):   in   year  6,   shoot  Cu   removals  by  OMDL  plants  varied  from  68  to  193  g  Cu/ha,  with  no  consistent  influence  of  genotypes.  For   the  OM2DL  plots,   the   genotype  did  not   influence   shoot  Cu   removal   and   its   values  were  similar  or  higher.  In  year  7,  shoot  length  and  DW  yield  were  higher  for  the  OM2DL  

plants  than  for  the  OMDL  ones,  which  promoted  shoot  Cu  removal.  This  was  sometime  more  marked  for  the  10-­‐8  variant.    ●  For  the  Linz-­‐Donawitz  slag  (LDS)  amended-­‐plots  in  year  6,  shoot  Cu  removal  peaked  with   the  variant  10-­‐8   in  one  plot,   reaching  254  g  Cu/ha.  For  all  genotypes,  differences  were   not   significant   between   the   P-­‐spiked   LDS   and  Unt   plots.   The   LDS   less   enhanced  shoot  DW  yield  than  two  compost  dressings  combined  with  dolomitic  limestone.  In  year  7,  tobacco  from  untreated  and  LDS-­‐amended  plots  had  again  a  lower  shoot  biomass  than  the   OM2DL   plants.   The   10-­‐8   variant   best   developed   in   some   cases,   but   the   genotype  influence  was  mostly  not  consistent.  Soil  amendment,  especially  the  second  dressing  of  compost  was  a  key  factor  to  maintain  the  shoot  yield.    Sunflower:  in  year  6  and  7,  leaf  chlorosis  occurred  on  many  sunflower  plants  growing  in  OMDL   amended   plots   with   high   total   soil   Cu   and   in   both   LDS-­‐treated   soils.   No  sunflowers  were  growing  in  the  UNT  soil.    ●   at   moderate   soil   Cu   contamination   (258   –   382   mg   Cu/kg):   in   year   6,   the   M2   mutant  performed  best  in  comparison  to  other  genotypes,  without  effect  of  the  second  compost  dressing.   The   shoot   DW   yield   of   the   M3   mutants   was   lower   than   that   of   motherline  plants  in  the  OM2DL  plots.  In  overall,  shoot  Cu  concentrations  were  similar  for  all  plants.  The  M2  mutant  showed  a  higher  shoot  Cu  removal  in  the  OM2DL  plot.  Shoot  Cu  removal  was   in   the   42   g   Cu/ha   range.   Due   to   lower   shoot   Cu   concentrations   than   in   previous  years,  this  was  lower  than  the  100  g  Cu/ha  reported  in  years  1  and  2.   In  year  7,  shoot  DW  yield  varied  from  0.5  to  36  t/ha  depending  on  plots  and  genotypes.    ●   at  high   soil  Cu   contamination   (894  –  1020  mg  Cu/kg):  plants  best  developed   in  years  6  and  7  in  all  OM2DL  plots  with  a  second  compost  dressing  in  year  6,  showing  the  need  to  maintain   soil   organic  matter.  A   single   compost  dressing   increased   the   shoot  DW  yield  more  than  the  incorporation  of  LD  slags.  The  Carmeuse-­‐LDS  and  P  spiked-­‐LDS  similarly  influenced  sunflower  growth,  and  both  M3  mutants  and  motherline  plants  died  in  these  plots  during  summer.  In  year  6  for  the  compost-­‐amended  plots,  M2  mutants  produced  a  higher   shoot   biomass   than   other   genotypes.   Both   mother-­‐line   plants   and   M3  mutant  developed  only  in  the  plots  with  recent  compost  dressing.  The  second  compost  dressing  decreased  shoot  Cu  concentrations  for  the  mutants.  Low  shoot  Cu  concentrations  (close  to   the   upper   critical   threshold   value   for   Cu   in   higher   plants)   for   the   OM2DL   plants  matched  with   their   high   shoot   biomass,   suggesting   a   dilution   effect.   In   overall,   shoot  biomass  is  a  main  driver  for  shoot  Cu  removal.  According  to  the  plant  density,  plots  and  sunflower  genotypes,  shoot  Cu  removal  varied  in  the  21-­‐105  g  Cu/ha  range.  ●   LDS   plots:   only   M1   and   M2   mutants   developed   on   these   plots.   Their   shoot   Cu  concentrations  were   higher   in   LDS   amended  plots   than   in   the  OM2DL  ones.   Shoot   Cu  removal  peaked  for  the  M1  mutant  mostly  due  to  their  higher  biomass.    In   year   7,   shoot   DW   yield   was   enhanced   in   the   OM2DL   and   the   LDS   amended   plots.  Sunflowers  did  not  grow  on  the  untreated  plots.  In  overall,  genotype  influence  was  not  consistent.  The  second  compost  dressing  in  year  6  was  the  key  factor  to  promote  both  shoot  yield  and  Cu  removal.  Between  years  4  and  6,  extractable  Cu  fraction  in  the  OMDL  plots  was   reduced   by   38%.   Since   the   experiment   started,   shoot   Cu   removal   fit  with   a  quadratic  function,  likely  following  reactions  of  compost  with  Cu,  nutrient  release  from  compost  decay  and  bioavailable  Cu  stripping.    Lommel  (BE):  Tobacco  clones  and  sunflower  mutants  were  cultivated  from  years  1  to  4.  Brassica  napus  and  Cannabis  sativa   (hemp)  were  also  implemented.   In  2013,  shoot  DW  yield   of   tobacco   was   similar   for   all   genotypes   ranging   from   1.3   to   1.7   t/ha.   For  sunflower,   shoot  DW  yield  varied   from  3.5   to  7.5   t/ha,  with   the  M1  mutant  producing  lower   shoot   biomass   than   the   control   plants   and   the   other  mutants.  Hemp  developed  

well  and   its  shoot  DW  yield  reached  17.5   t  ha-­‐1  yr-­‐1.  Shoot  metal   removals   for   tobacco  were   in   the   following   ranges   (in   g/ha/yr):   Cd   14   –   20,   Pb   8.6-­‐11.9,   and   Zn   252-­‐331.  Genotype  had  an  influence  (e.g.  higher  shoot  removal  for  Cd/10-­‐8  variant  and  Zn/7-­‐19  variant).  Compared   to   tobacco,   shoot  metal   removals  of   sunflower  were  higher   for  Zn  (1992-­‐2504)  and  slightly  higher  for  Cd  (17.7-­‐23.6)  and  Pb  (17.7-­‐42.5),  with  a  genotype  influence.   The   phytoextraction   by   hemp  was   7   g   Cd,   41   g   Pb,   and   1355   g   Zn/ha.   The  phytoextraction  of  Cd,  Pb  and  Zn  by  all   tested   sunflowers  was  higher   than   that  of   the  tobacco   clones   and  hemp.  Tobacco   clones  had  higher   shoot   Cd   and  Pb   concentrations  but  the  higher  shoot  DW  yield  of  sunflowers  lead  to  higher  shoot  metal  removals.  Hemp  production  on  metal-­‐contaminated  soil  could  be  relevant  if  cutting  the  pollutant  linkage  to  food  along  with  an  economic  profit  from  the  plant-­‐based  feedstock  (e.g.  fibre)  is  the  primary   goal   instead   of   the   other   ecosystem   services   (e.g.   decreases   of   labile   metal  pools,  decontamination,  and  soil  remediation).    In  2014,  the  biomass  production,  and  consequently  the  shoot  metal  removal,  was  higher  than  in  2013,  especially  for  tobacco  (tobacco  3.6-­‐4.9  t/ha;  sunflower  5.8-­‐9.5  t/ha).  Shoot  removals  were  59.6-­‐122  g  Cd,  38-­‐70  g  Pb  and  1027-­‐1926  g  Zn/ha  for  tobacco,  32-­‐61  g  Cd,  17.5-­‐34.5  g  Pb,  and  2624-­‐5745  g  Zn/ha  for  sunflower.  This  highlights  the  influence  of  annual  climatic  conditions  on  shoot  metal  removals.    Bettwiesen   (CH):   The   crop   rotation   is   based   on   four   sunflower   and   five   tobacco  genotypes   with   higher   metal   tolerance   and   accumulation   properties   for   stripping  bioavailable   Zn   and   Cd   excess   in   topsoil.   After   5   years,   the   labile   Zn   pool   in   soil   was  lowered  by  45-­‐70%.  A  Mass  Balance  Analysis  confirmed  soil  Zn  decontamination  in  line  with  plant  Zn  uptake.  The  plants  partially  take  Zn  from  the  non-­‐labile  pool  of  the  total.  The   ‘stripping’   of   bioavailable   Zn   is   feasible   within   a   few   years   period.   To   decrease  available  soil  Zn  below  the  Swiss  threshold  value,  the  phytomanagement  would  take  3-­‐12  years  at  moderate  available  Zn  levels  and  5-­‐25  at  high  levels.  Various  plant  densities  and  intercropping  of  sunflower  with  tobacco  in  early  spring  are  further  explored.    3.1.3.2.3.2.  (aided)  phytoextraction  using  woody  SRC    The  capacity  of  poplar  and  willow  to  colonize  hostile  environments  such  as  mine  wastes  is  recognized.  Numerous  Salix  and  Populus  clones  have  been  screened,  and  show  great  variation   in  biomass  production,  TE   tolerance  and  accumulation  patterns   in   roots  and  leaves   between   clones.   Most   promising   poplars   and   willows   (according   to   climatic  conditions   and   shoot   TE   concentrations)   were   assessed   at   6   sites:   Lommel   (HAU),  Högbytorp  and  French  trial  (SLU),  Freiberg  (LfULG),  Piedrafita  (ES)  and  Phytagglo  (FR).      Freiberg,   Halsbrücke   Krummenhennersdorf   (DE):   this   9-­‐old   field   trial   is   a   SRC  plantation  on  contaminated  agricultural   land.  Shoot  DW  yields   reached  15   t/ha/yr   for  poplar  SRC  and  14-­‐19  t/ha/yr  for  willow  SRC,  corresponding  to  common  values  (6-­‐16  t  DM   ha-­‐1   yr-­‐1).   Stem   wood   and   bark   ionomes   of   poplars   and   willows   depended   on  genotypes,  particularly  for  Cd  (3-­‐10  mg/kg  in  wood;  8-­‐30  mg/kg  in  leaves),  and  clonal  differences   were   higher   across   willows.   Values   were   higher   in   willows   compared   to  poplars.  Willow  cultivars  Tora,  Tordis  and  Gudrun  displayed  the  highest  wood  and  foliar  Cd  concentrations  among  all  cultivated  clones  of  poplars  and  willows.  The  ratio  of  the  3-­‐year-­‐average   concentration   in   leaves   compared   to  wood  was  about  2.6  and  2.7   for  Cd  and  about  5.7  and  6.2  for  Zn  in  poplar  and  willow,  respectively.  Bark  Cd  concentrations  account   for   triple   (poplars)   to   fourfold   (willows)   of   those   in   stem   wood   while   Pb-­‐concentrations  did  not  differ  between  the  compartments.  Foliar  As-­‐concentrations  were  

mostly  below  the  detection  limit.  Wood  As  concentrations  varied  from  2  to  4  mg/kg.  For  the   third   rotation,   willow   Tora   produced   the   highest   biomass   out   of   all   poplars   and  willows   (followed  by  Tordis)   and   it   displayed   relatively   high  Cd   and  Pb   accumulation  capacity.  This   confirmed  Tora   as   a   relevant   choice   for  metal   (Zn,   Cd)  phytoextraction,  and  Tordis  as  well  (for  Cd).    Piedrafita   do   Cebreiro   (ES):   plots   with   Salix   smithiana   and   S.   atrocinerea,   in  monoculture  and  inter-­‐cropped  with  Alnus  glutinosa,  and  Salix  cv.  Tora  were  established  in   autumn   2012.   Plots   were   unamended   or   amended  with   5%   (w/w)   compost.   Plant  survival   was   higher   for   S.   smithiana   than   either   S.   atrocinerea   or   Tora.   Plant   growth  (height/spread)  and  leaf/stem  Cd/Zn  concentrations  were  recorded  in  2013  and  2014.  No   significant   effects   of   inter-­‐cropping   on   plant   growth   have   been   recorded   to   date  (although  there  was  a  tendency  to  increased  tree  height  in  intercropped  plots  in  2012).  Intercropped  S.  smithiana  plots  showed  higher   leaf  Cd  and  Zn  concentrations  (approx..  12  and  1400  mg/kg).  Changes  in  total  or  NH4Cl-­‐extractable  concentrations  of  Cd  and  Zn  are  not  yet  consistent.      Changes  in  soil  exposure:  At  Freiberg,  generally  total  soil  Cd  decreased  (e.g.  3.4  to  2.9  mg/kg  for  Weser  6  poplar;  2.5  to  1.0   for  Tora  willow)  between  2011  (year  6)  to  2013  (year   8).   The   rhizosphere   of   willow   and   poplar   clones   at   contaminated   and   adjacent  reference   sites   showed   a   lower   pH   for   SRC   compared   to   arable   land   (with   a   higher  decrease  for  willow  SRC).  Since  the  SRC  implementation  in  2005,  the  initial  soil  pH  of  5.7  (CaCl2)  dropped  in  average  to  5.2.  However,  soil  pH  increased  in  poplar  plots  after  the  harvest   in   year  8,  whereas   a   further  decrease  occurred   for  willow  plots   especially   for  Tordis  cultivar.  NH4NO3-­‐extractable  soil  Cd  and  Pb  were  roughly  10   fold  higher  under  SRC   and   grassland   compared   to   annual   cropped   land.   In   contrast,   mobile   As   in   soil  decreased   for   SRC   (3   to   9   fold)   compared   to   annual   cropped   land   (lowest   values   for  willow   SRC).   Soil   pH   was   higher   under   winter   wheat   with   the   lowest   NH4NO3-­‐extractable  soil  Cd  in  year  2.  NH4NO3-­‐extractable  soil  Cd  decreased  for  Jorr,  Sven,  Tora  willow  SRC  and  Max  3  poplar  SRC.    Soil  microbial  communities:  At  Freiberg,   arylesterase  and  arylsulfatase  activities  did  not  differ  between  SRC  and  control  plots.  Alkaline  phosphatase  was  higher  in  SRC  plot.  At  Piedrafita,   the   phytomanagement   (willow   SRC)   induced   changes   in   soil   microbial  communities  showing  influence  of  compost  and  the  vegetation  covers,  notably  between  mono-­‐  and  co-­‐cultivated  S.  smithiana.  The  diversity  of  bacterial  communities  decreased  with   time,   with   higher   value   in   phytomanaged   soils,   but   without   clear   plant-­‐induced  effect.  Most   soil   enzyme   activities   increased   in   the   phytomanaged   soils,   with   a   plant-­‐induced  effect  more  pronounced  in  the  compost-­‐amended  soils.  No  clear  effect  of  the  co-­‐cropping  of  S.  smithiana  with  Alnus  glutinosa  was  evidenced.  At  Phytagglo,  willows  (S.  viminalis)  were  planted  on  an  alkaline  technosol  developed  on  dredged   sediments   contaminated   by   Zn,   Pb,   Cu,   and   Cd.   Citric   acid   based   product,  ferrous   sulfate,   and   elemental   S   were   separately   incorporated   for   investigating   the  effects   of   soil   acidification.   In   year  1,   extractable  metal   fraction  decreased   for  Pb,   and  remained   steady   for   Cu,   Cd   and   Zn,   and   the   survival   rate   of   willows  was   90%.   Their  foliar  Cd  and  Zn  concentrations  ranged  between  1.5-­‐3.8  and  233-­‐1176  mg  kg-­‐1.  Foliar  Cu  concentrations  (7-­‐12  mg  kg-­‐1)  were  similar  to  common  values  of  willows.    Högbytorp  (landfill  leachate  trial,  SE):  For  14  commercial  willow  SRC  plantations  long-­‐term  grown  (ca.  15  years)  on  agricultural  soil  in  Sweden,  total  topsoil  Cd  decreased  (ca.  13%  on  average)  compared  to  adjacent   fields  cultivated  with  cereals   in  common  crop-­‐

rotations.  The  biomass  productions  on  these  SRC  fields  were   lower  than  the   indicative  10   t   DM   ha-­‐1   yr-­‐1   expected   nowadays   in   well-­‐managed   fields.   Farmers   had   lack   of  experience  in  growing  such  crops,  and  beneficial  incentives  in  terms  of  subsidies  caused  limited   engagement   throughout   the   process.   Here,   treatments   consist   of   three   supply  rates   of   landfill   leachate   (irrigation   started   in  2005  and   carried  out  until   2010)   and   a  control,  with  two  willow  clones,  i.e.  Tora  (Salix  schwerinii  x  viminalis)  and  Gudrun  (Salix  dasyclados  variety  with  partly  Russian  origin,  more  frost-­‐tolerant  than  Tora).    For  willows,  shoot  concentrations  varied  in  the  1-­‐4.5  mg  Cd  and  40-­‐120  mg  Zn  kg-­‐1  DW  ranges.  Tora  showed  higher  shoot  Cd,  Co,  Mn,  Pb  and  Zn  concentrations  than  Gudrun  for  leachate   irrigated-­‐plots.   Tora   had   higher   shoot   Cd,   Co   and   Zn   concentrations   for   the  plots  irrigated  at  the  second  supply  rate  compared  to  other  treatments  and  the  control.  Gudrun   displayed   higher   shoot   concentrations   for   all  metals   in   the   control   plots.   The  leachate  treatments  did  not  influence  the  shoot  Cr  and  Cu  concentrations  in  both  clones,  even  though  these  concentrations  were   low  for  Tora  on  the  control  plots  compared  to  the   treatments   1-­‐3.   Nickel   was   the   only   metal   with   higher   shoot   concentrations   in  Gudrun   for   all   treatments.   Total   shoot   N   concentration   was   roughly   similar   for   both  clones,  except  a  higher  concentration  for  Gudrun  on  the  control  plots.    Lommel  (BE):  Willow  clones  ‘Belgisch  Rood  (BR)’  (Salix  x  rubens  var.  basfordiana)  and  ‘Tora’  (Salix  schwerinii  x  Salix  viminalis)  were  compared.  Shoot  DW  yields  after  3  years  were  5.4  (BR)  and  9.0  (Tora)  t/ha.  Again  shoot  Cd  concentrations  were  higher   in  Tora  (30  mg/kg)  than  in  BR  (24  mg/kg)  and  shoot  Zn  concentration  as  well  (1268  and  918  mg   Zn/kg).   The   Tora  willow   clone   had   a   Cd   and   Zn   removal   capacity   (274   g   Cd   and  11  417g  Zn  ha-­‐1)  which  is  twice  as  high  as  that  for  BR  clones.  Both,  the  higher  biomass  production  (ton  ha-­‐1)  and  metal  uptake  capacity  (mg  kg-­‐1  DW)  make  the  TO  willow  clone  the  favorable  clone  to  select  for  phytoextraction  applications.    1.3.3.2.   TE   hyperaccumulators:   can   accumulate   high   concentrations   of   metal(loid)s  (e.g.  Cd,  Ni,  Zn,  Se,  and  As)  in  their  above-­‐ground  biomass  and  possess  some  economic  added   value   (renewable   biomass   for   bio-­‐economy   and   bio-­‐ores).   Variations   in   both  biomass   production   and   TE   accumulation   within   populations   of   hyperaccumulators,  such  as  Noccaea  caerulescens  (for  Zn/ecocatalysis),  Alyssum  murale,  A.  bertolonii  and  A.  corsicum,   (for   Ni   phytomining)   allows   for   the   selection   and   breeding   of   improved  phytoextractor  plants.  The  main  bottleneck  limiting  their  practical  application  is  the  low  biomass   production   of   most   species   (except   some   Ni-­‐hyperaccumulators)   and   the  number   of   cropping   cycles   required   for   clean-­‐up.   However   this   number   is   generally  reduced  when  the  option  of  bioavailable  TE  stripping  is  considered.    

At  Piedrafita,  the  Cd/Zn  hyperaccumulator  N.  caerulescens  and  its  inter-­‐croppings  with  Lupinus  albus  and  Lotus  corniculatus  were  assessed.  Inter-­‐cropping  N.  caerulescens  with  the  legume  L.  corniculatus  tended  to  increase  Cd  accumulation  by  the  hyperaccumulator.  Other  Lotus   species  show  potential   for   incorporating   into  GRO  due   to   their  worldwide  distribution  and  high  adaptation  to  a  number  of  abiotic  stresses.  Candidates  with  good  potential   for   cultivation   in   degraded   or   marginal   soils   include   L.   corniculatus,   L.  uliginosus,  L.  tenuis  and  L.  creticus.  The  soils  planted  with  N.  caerulescens  and  those  co-­‐cropped   with   L.   corniculatus   have   similar   soil   microbial   communities.   Compost   and  plants   induced   changes   in   soil  microbial   communities.  Mono-­‐cultures   and   co-­‐cultures  tend  to  separate  over  time.  The  diversity  of  bacterial  communities  increased  with  time,  with  a  higher  diversity  in  phytomanaged  soils  compared  to  untreated  soils,  and  a  higher  diversity  in  planted  soils  compared  to  unplanted  soils.  

At  Phytagglo,  seven  plots  were  set  up  with  potentially  acidifying  properties,  i.e.  legume  plant  (Lupinus  albus),  a  citric  acid  based  product,  peat-­‐like,  ferrous  sulfate  and  elemental  sulphur,  Arabidopsis  halleri  and  control.    At   Reppel,   since   2004,   Pteris   vittata   L.,   an   As   hyperaccumulator   was   cultivated   for  bioavailable  As   stripping.  Generally,   frond  DW  yield  was  doubled   in   the   contaminated  soils  compared  to  the  uncontaminated  control  soil.  Soil  treatments,  i.e.  Beringite  (B,  5%  w/w),   iron  grit  (Z,  1%  w/w)  and  their  combination  (BZ),  and  season  did  not   influence  annual   frond   yield,   except   differences   between   B   and   BZ   in   November   and   between  November   and   May   for   the   untreated   (Unt)   and   B   soils.   On   the   2006-­‐2013   period,  leachate  As  concentration  remained  lower  in  Z-­‐treated  soils  than  in  the  Unt  and  B  soils.  Mean  values  of  frond  As  concentrations  (in  mg  As/kg)  varied  in  the  60-­‐171  range  for  the  control   soil   and   in   the   970-­‐2870   range   for   the   contaminated   soils.   Frond   As   removal  varied  from  3.89  to  2.28  g  As/m²  in  the  decreasing  order:  Unt,  B  >  BZ,  Z.        3.2  Valorization  of  plant  biomass  produced  on  TE-­‐contaminated  sites    As   a   result   of   plant   and   culture   management,   Gentle   Remediation   Options   (GRO)  produce  plant  biomass  (herbs  or  woody  biomass).  Depending  on  the  GRO  set  up  on  the  polluted   site   and   the   type   of   plant   used,   harvested   plant   parts   may   contain  concentrations  of  TE  that  may  be  higher  than  those   found   in  similar  vegetation  grown  on  uncontaminated  soils.  This  is,  in  particular,  the  case  of  phytoextraction,  which  leads  to  metal-­‐enriched  plant  biomass.  These  plants  may  enter  valuation  pathways  if  (i)  TE  do  not  disturb  the  functioning  and  the  performance  of  the  process,  (ii)  if  the  TE  transfer  is  controlled   and   (iii)   if   such   plant   use   complies   with   current   regulation.   To   our  knowledge,  by  far,  plant  biomass  on  contaminated  lands  was  only  produced  for  scientific  purpose   to   be   used   in   demonstration   projects   such   as   GREENLAND.   As   a   potential  advantage,   these   plants   will   not   compete   with   plants   grown   on   agricultural   lands   as  contaminated  lands  are  not  suitable  for  food  production.  On  contaminated  lands,  plants  may   serve   to   provide   feedstocks   and   non-­‐food   products   for   bioenergy   and,   thus,  may  contribute  to  achieve  the  EU  aim  by  2020,  i.e.  to  get  20%  of  its  energy  from  renewable  sources.    In   GREENLAND,   our   approach  was   to   select   routine   pathways   for   plant   biomass   as   a  basis   to  discuss   the  possible  advantages  and  potential   limitations,   regarding   technical,  social   and   regulatory  aspects,   of  using  plant  biomass  produced   from  TE  contaminated  soil  into  these  pathways.  In  addition,  three  emerging  processing  pathways  were  selected  and  discussed  based  on  existing  knowledge.  Thus,  combustion  and  anaerobic  digestion  were  selected  as  established  pathways  whereas  solvolysis,  flash  and  slow  pyrolysis  were  selected   as   emerging   technologies.   Technical   assessment   was   based   on   assays.   They  were   performed   with   plants   cultivated   for   the   purpose   of   phytoextraction   leading   to  metal-­‐enriched   biomass.   All   plants   used   in   assays   were   provided   by   GREENLAND  partners   who   owned   field   sites.   Assays   were   performed   with   equipments   owned   by  GREENLAND  partners.  Table  3.2.1  details  the  processes  and  plants  used  in  assays.    Table  3.2.1:  Type  of  process  and  plant  used  in  assays.  Process   Test  scale   Plant   Targeted  

metal  Combustion   Pilot  (40kW)   Willow  ‘Tora’   Zn,  Cd       Poplar  ‘Max3’   Zn,  Cd  

    Mix  willow,  poplar   Zn,  Cd  Anaerobic  digestion   Laboratory    

(5L  reactor)    Sunflower   Zn  

Solvolysis   Laboratory    (110cm3  reactor)  

Tobacco   Zn,  Cd  Cu  

Flash  pyrolysis   Laboratory    (100g  reactor)  

Willow  Sunflower  

Zn,  Cd  Zn  

    Tobacco     Zn,  Cd  Cu    

Slow  pyrolysis   Laboratory    (100g  reactor)  

Tobacco   Zn,  Cd  

 Acceptance   and   feasibility   assessment   were   realized   for   combustion   and   anaerobic  digestion   based   on   interviews   of   installation   operators   in   several   European   countries  (France,   Austria,   Germany,   Sweden).   Regarding   regulatory   aspects,   the   assessment  consisted   in   a   review   of   current   European   regulation   and   examples   of   national  regulations   related   to   combustion   and   anaerobic   digestion   focused   on   plant   biomass  utilization.   This   review   was   the   basis   to   discuss   possibilities   to   use   plant   biomass  produced  on  TE  contaminated  lands  in  these  processes.    KEY  RESULTS    I   Assays  were  performed   to  determine   the   fate  of   the  TE   in   the   resulting  products  of  each  conversion  process.  Combustion,  defined  as   thermochemical  conversion  of  biomass,  occurs   in  combustion  plants  or  boilers  in  which  fuels  are  oxidized  to  use  the  heat  generated.  For  all  assays,  Zn  occurred   mainly   in   the   fly   ashes.   The   bottom   ashes   represented   the   second  compartment   for   the   occurrence   of   Zn  whereas   the   gaseous   fraction   of   the   flue   gases  represented  a  minor  compartment  for  Zn  emissions.  The  distribution  was  not  depending  on   the   initial   burnt   wood,   i.e.   virgin   wood   (control)   or   Zn   enriched   wood  (phytoextraction).  Similar   results  have  been   found   for  Cd.   Independently  of   regulation  issues,  assays  allowed  to  conclude  that  the  burning  of  plant  biomass  naturally  enriched  with  metals   in   industrial   or   collective   boilers   could   be   possible,   as   they   are   normally  equipped   with   efficient   systems   to   reduce   dust   emissions.   Depending   on   the   TE  concentration   in   bottom   ashes   and   national   legal   framework,   bottom   ashes   could   be  valued  by  land  spreading.  Concerning  fly  ashes,  the  results  invite  to  perform  further  in-­‐depth   analysis   of   current   practices   regarding   separation   of   ashes   and   valorisation  pathways.    Anaerobic  digestion,  a  biological  process  performed  by  the  combined  action  of  several  micro-­‐organisms   in   the   absence   of   oxygen,   ends   up   in   partial   degradation   of   organic  matter  and   leads   to   formation  of  biogas  and  digestate.  Medium  Zn-­‐enriched  sunflower  showed   similar   biogas   composition   as   typical   biogas.   This   result   evidenced   that   the  presence   of   Zn   in   sunflower   did   not   modify   the   composition   of   biogas.   Results   also  showed  that  Zn  did  not  inhibit  biogas  production.  Due  to  technical  problems,  the  assay  performed  on  high  Zn-­‐enriched  sunflower  was  not  conclusive.  Nevertheless,  during  the  biogas  monitoring  which   lasted  10  days,  we  could  observe  that  biogas  production  was  not   inhibited.   As   expected,   Zn   was   measured   in   digestates.     Indeed,   at   55°C,   the  temperature  of    the  anaerobic  digestion,  no  Zn  volatilization  can  occur.  Depending  on  TE  

concentration   in   digestates   and   legal   framework,   digestates   could   be   valued   by   land  spreading  or  by  composting.    Solvolysis,   chemical   decomposition   of   biomass   with   a   solvent   under   pressure,  investigated  metal  behaviour  in  biomass  converted  by  sub-­‐  and  supercritical  conditions.  Cu  was  mainly  found  in  the  liquid  phase  during  the  heating  step  or  in  the  residual  solid,  depending   on   the   temperature.   Zn   was   mainly   found   in   the   liquid   phase   during   the  heating   step   whereas   Cd   was   mainly   found   in   the   residual   solid.   Carbon   is   almost  exclusively   found   in   the   residual   solid.   Some  molecules   of   interest   for   fine   chemistry  were  found  in  the  liquid  phase  but  in  very  small  amount  which  not  allowed  to  quantify  them.  In  the  solid  residues,  Cu  concentrations  were  too  high  to  consider  the  usage  of  the  solid   residue   as   an   organic   amendment.   The   idea   was   then   to   use   the   solid   phase  enriched  with  metals  as  raw  material  to  produce  polymetallic  catalysts  which  could  be  used   in   industrial   biotechnologies   and   chemocatalytic   processes.   Preliminary   assays  showed   that   the  metal   concentrations  were   too   low   to   evidence   a   catalytic   activity   of  these  residues.  Solvolysis  can  be  used  as  a  pre-­‐treatment  leading  to  a  significant  reduced  biomass  and  a   liquid.   From   the  view-­‐point  of   the   industry,   it  would  be  easier   to   get   a  metal  free  liquid  phase  that  could  be  rejected  in  nature.    

   Flash  and  slow  pyrolysis.  Pyrolysis  is  the  thermochemical  decomposition  of  (biomass)  material  at  moderate  temperature  and  in  oxygen  deficient  conditions  resulting  in  3  end  products:  char,  oil  and  gas.  Flash  pyrolysis  typically  uses  moderate  temperatures  (450  –  600°C),  a  very  high  heating  rate  and  a  very  short  vapor  residence  time.  Flash  pyrolysis  targets   the  pyrolysis   liquid  as  end  product.  Low  process   temperature  and   long  vapour  residence  times  ("slow"  pyrolysis)  favour  the  production  of  char.  In  flash  pyrolysis,  the  Cd  concentrations  in  the  aqueous  fractions  were  never  higher  than  12.3%  of  the  %wt  of  Cd  present  in  the  original  biomass.  The  recovery  of  Zn  in  the  aqueous  fraction  is  much  lower  and  did  not  exceed  2.8%  of  the  %wt  of  Zn  present  in  the  biomass.  Cu  content  in  the  aqueous  pyrolysis  oil  after  flash  pyrolysis  of  the  Cu-­‐rich  biomass  was  relatively  low.  The   tar   fractions   of   tobacco   and   sunflower   contained   in   all   cases  more   target  metals  than   the   corresponding   aqueous   fractions.   To   be   used   as   a   renewable   fuel,   the  physicochemical   properties   of   the   liquid  must   be   investigated   as  well   as   the  potential  impact  and  constraints  associated  to  the  presence  of  the  metals  for  this  usage.  Further  research   efforts   are   needed   to   investigate   these   points.   Metal   enriched   biomass   was  successfully  valorized  by  slow  pyrolysis  and  subsequent  physical  activation  by  steam  in  products  with  added  value,  in  particular  low  cost  activated  carbons.      From   safety   point   of   view,   considering   previous   cited   valuation   pathways   of   metal  enriched   biomass,   we   were   not   able   to   identify   any   major   reason   to   stop   further  consideration  of  any  of  those  routes  for  safety  reasons.    II  Interviews    Operators   of   anaerobic   digestion   (AD)   platforms   and   actors   of   the   wood   bioenergy  sector   were   interviewed   to   assess   the   potential   acceptance   of   using   plant   biomass  produced  by  GRO  on  metal  contaminated   lands   in  their   installations  and  network.  The  reasons   of   acceptance   or   not  were   investigated   by   separating   phytostabilisation   from  phytoextraction.  Selection  of  AD  platforms  and  boiler  operators/owners  was  based  on  countries  among  those   represented   in   GREENLAND   which   used   wood   and   energy   crops   as  

fuel/feedstocks  at  a   significant   rate   in   combustion  and  AD.  As  a   result,  8  actors  of   the  wood  energy  sector  from  France,  Germany  and  Sweden,  and  11  AD  platforms  operators  from   France,   Germany   and   Austria   were   interviewed.   The   questionnaire   was   asking  about   installation   characteristics,   plant   characteristics,   performed   analyses   and  phytotechnologies.  Results  from  questionnaires  suggested  that  plant  biomass  from  phytotechnologies  could  be  used   in  AD  and   combustion,   under   conditions.   From   the  view-­‐point  of   interviewed  actors,   main   limitations   related   to   additional   controls   in   process   end-­‐products   and  installations   that   might   generate   additionnal   costs.   In   most   cases,   price   of  phytotechnologies   biomass  was  mentionned   as   a   driver   to   potentialy   use   plants   from  metal  contaminated  soils.  It  should  be  similar  to  market  price  for  feedstocks  and  fuels,  less  expensive  or  free.  Plants  used  in  phytostabilisation  or  phytoexclusion  were  thought  to  be   less   risky  and,   consequently,  benefited   from  a  better   theoritical   acceptance   than  those  issued  from  phytoextraction.    III  Regulation    The   classification   of   the   plant   biomass   produced   on   contaminated   land   (biomass   or  waste?)   is  essential   to  choose   the  appropriate  valuation  pathway,  and   thus,  assess   the  profitability  or  the  cost  due  to  gentle  remediation  options.  By  far,  this  question  is  solved  neither   at   the   European   level   nor   at   the   local/national   level.   To   know   how  European  regulators   would   consider   biomass   produced   on   contaminated   soils   by  phytotechnologies,  we  asked  some  of  them  through  the  GREENLAND  advisory  board.  A  first  comment  from  regulators  related  to  the  fact  that  this  point  was  never  discussed  yet.  One   reason   could   be   that   the   amount   of   plants   produced   on   contaminated   sites   for  remediation  purpose  or   for  bioenergy  production   is  by   far  not  significant,  as   it   is  only  produced   for   scientific   purpose.   Except   one   regulator   who   had   the   feeling   that   this  biomass  could  be  classified  as  agricultural  product,  other  regulators  had  the  tendency  to  classify  the  biomass  from  phytotechnologies  as  waste.  These  answers  could  be  related  to  the   way   of   the   questions   were   formulated   that   could   have   orientated   the   regulator  answers.   Nevertheless,   these   answers   highlighted   the   fact   that   it   is   necessary   to  evidence   harmlessness   of   metal-­‐enriched   biomass,   bring   information   on   TE   transfer  control  to  regulators  and  clarify  product  vs  waste  consideration.    A   state   of   the   art   on   European   and   national   regulations   was   performed   related   to  combustion  aspects,  as  this  valuation  pathway  is  the  most  important  energy  conversion  route   for  biomass  produced  on  uncontaminated  soils,  and  anaerobic  digestion  aspects,  highlighting  metal  emission  limit  values  or  metal  input  fuel  concentrations.  Finally,  plant  biomass   from  phytotechnologies   are  not   specifically   addressed   in   regulation   (national  and   European   level).   Results   of   the   regulation   study   and   the   combustion   assays  performed  on  plant  biomass  used  in  phytoextraction  highlighted  the  need  to  separate  fly  ashes  from  bottom  ashes  (in  countries  where  it   is  not  already  done)  to  valuate  bottom  ashes  more  easily  and  to  manage  fly  ashes  accordingly  to  their  TE  content.  Results  and  regulation  interpretation  with  less  metal  enriched  plant  biomass,  i.e.  phytostabilising  or  phytoexclusing  plants,  could  obviously  be  different.  As  Zn  and  Cd  are  mostly  recovered  in  fly  ashes  than  bottom  ashes,  it  should  be  possible  to  land  spread  more  easily  bottom  ashes   than   fly   ashes.   Nevertheless,   a   low   concentration   in   plant   biomass   doesn’t   not  imply  that  ashes  resulting  from  this  plant  can  be  spread  on  land.  Further  research  might  focus  on  combustion  assays  with  phytostabilising  plants  to  answer  more  precisely  this  point.    

 3.3  Harmonisation  of  methods  to  assess  the  bioavailability  of  TE  and  development  of  tool  set  to  monitor  the  sustainability  of  GRO    Objectives  For  the  GRO  options  to  be  accepted  by  decision  makers,   the  methods  for  evaluation  of  GRO  success  should  be  widely  available,  provide  robust  results  and  should  be  suitable  for  monitoring  of  soil  health  and  sustainability.  For  this,  we  need  to  provide  proof  of  the  suitability   of   GROs   for   risk   reduction   of   contaminated   sites   by   collecting   comparable  results   using   the   same  methods   throughout   Europe.   Hence,   the   objectives   of   this  WP  were:  i)  to  select  methods  to  be  used  as  indicators  for  GRO  success  and  as  sustainability  monitoring   tools;   and   ii)   to   select/harmonise   methods   describing   the  bioavailable/bioaccessible  TE  fractions  among  European  case  studies.      Methods  Two   sets   of   tests   (so   called   test   batteries)   were   pre-­‐selected   based   on   available  literature  and  the  experience  of  the  consortium;  a  chemical  one  to  quantify  TE  exposure  in   untreated   soils   and   GRO-­‐managed   soils,   and   a   biological   one   to   characterize   soil  ecotoxicity   and   functionality.   The   chemical   tests   included   extractions  with  aqua  regia,  0.05  M  EDTA,  1  M  NH4NO3,  0.1  M  NaNO3  and  H2O.  The  biological   test  battery   included  ecotoxicity   tests   (plantox  with  dwarf  beans,   lettuce  and   turnip;   activity  of  plant   stress  enzymes,   and   soil   invertebrates,   such   as   earthworms   and   nematodes);   specific   soil  biochemical  functions  and  a  number  of  tests  with  soil  microorganisms.    The  common  sampling  strategy  was  developed  for  collection  of  soil  from  the  test  fields.  Soil  was  sampled   following   the  agreed  procedure   from  the  case  sites   representing   the  main   GROs:   phytoextraction   in   Belgium,   Sweden,   Germany   and   Switzerland;   aided  phytoextraction   in   France   and   Spain;   and   aided   phytostabilisation   or   in   situ  stabilization/phytoexclusion  in  Poland,  France,  Spain  and  Austria.    All  partners  managing  case  sites  applied  the  chemical  test  battery  on  their  own  soils  two  times  during  the  four  years:  one  run  during  the  first  year  and  the  second  run  during  the  third   year   of   the   project.   At   least   one   sample   from   untreated   and   treated   plots   per  treatment  method  was  analyzed  in  at  least  3  replicates.    For   the   implementation   of   the   biological   test   battery,   the   sub-­‐samples   were   sent   for  analysis   to   HAU,   INRA   and   INERIS   (ecotoxicity   tests).   These   tests   were   implemented  twice,   the   first   and   the   third   year.  Methods   that   are  more   demanding   and   require   an  advanced   analytical   base,   samples  were   sent   to   selected   consortium   laboratories   that  are  able  of  performing  those  tests  (HAU,  UF,  CSIC).  Only  one  measurement  occasion  was  applied  (the  first  year  of  the  project).      Results  The   chemical   test   battery   showed   that   extractable   trace   element   (TE)   concentrations  generally   decreased   more   significantly   in   soils   managed   by   in   situ   stabilisation  combined  with  phytoexclusion,  phytostabilisation  or  phytoextraction  than  in  soils  only  managed  with  phytoextraction.    The  extractant  strength  towards  dissolution  of  several  tested  TE  (e.g.  Cd  and  Zn)  was  in  increasing  order:  H2O  <NaNO3  <NH4NO3  <EDTA  <aqua  regia,  despite  the  broad  range  of  the   total   soil   Cd   and   Zn   concentrations.   Good   result   reproducibility   was   achieved   as  similar   results  were  obtained  by  repeating   the  extractions  on   the  samples  collected  at  

the   same   places   of   each   site   two   years   later.   Salt   solutions   were   less   effective   in  extracting   oxyanions,   such   as   arsenic   (As),   than   distilled   water.   Extractions   that   best  describe  phyto-­‐  and  bioavailable  As  should  be  further  investigated.        Among   the   chemical   extractions,   the   NH4NO3   and   EDTA-­‐extractions   showed   the  most  frequent   differences   in   the   extracted   TE   concentrations   between   the   treated   and  untreated   soils,   while   the   most   frequent   correlations   with   the   biological   responses  occurred  for  NH4NO3,  followed  by  NaNO3-­‐extractable  TE  pools.  Pseudo-­‐total  (aqua  regia  extractable)   concentrations   showed   no   significant   correlation   with   the   biological  responses.  Among  the  bioindicators  (plants,  earthworms  and  nematodes),  dwarf  beans,  especially  through  root  mass,  followed  by  shoot  length,  and  stress  enzyme  activities,  were  the  most  responsive   indicators   to   the   soil   treatments.   Even   though   the   selective   chemical  extractions   did   not   always   show   statistically   significant   changes   in   TE   extractability,  dwarf   beans   and   stress   enzymes   developed   a   stronger   response   to   the   tested   GRO  options.   Generally,   the   plant   growth   decreased   with   higher   extractable   TE  concentrations   in   soil,   while   bean   stress   enzymes   reacted   in   the   opposite   way,   i.e.  increased  with  increasing  TE  extractability.    The  soil  biochemical  properties  positively  responded  to  the  different  phytoremediation  options,   with   arylesterase,   urease   and   protease   enzymatic   activities,   nitrification   and  ammonification   potentials   responding   in   all   studied   cases.   The   β-­‐glucosidase   activity  only  responded  in  sites  where  organic  amendments  were  used.  Similar  responses  were  observed   for   soil   N   and   P,   which   significantly   changed   only   in   soils   amended   with  organic  matter.    The   measured   microbiological   and   biochemical   endpoints   indicated   significant  improvement   of   soil   functionality   in   soils  with   the   heaviest   contamination   and  where  organic  amendments  were  used  such  as  in  the  French,  German  and  Swiss  sites.  In  soils  with  moderate  contamination  such  as  those  of  Belgian  and  Swedish  sites,  only  some  of  the   microbiological   and   biochemical   endpoints   were   improved   by   the   adopted   GRO  strategy.    Among   the   microbiological   endpoints,   the   most   responsive   were   nitrification   and  ammonification  potentials,   followed  by  soil  enzymatic  activities,  which  were  similar   to  soil  respiration,  and  the  least  responsive  was  microbial  biomass.  These  microbiological  and   biochemical   endpoints   are   in   a   good   agreement   with   soil   toxicity   data   and   TE  solubility   and   mobility   estimated   by   chemical   extractions,   and   therefore   are   robust  indicators   on   which   management   decisions   can   be   based   on.   It   is   expected   that   soil  microbial   biomass   and   functional   activity   increases   during   the   early   stages   of  phytomanagement  TE  contaminated  sites  up  to  typical  levels.  This  is  because  soils,  like  any  other  ecosystem,  have  an  own  maximum  carrying  capacity,  which  is  site  specific  and  depends  on  the  pedo-­‐climatic  conditions,  vegetation  and  site  management.      Functional  gene  diversity  of  the  soil  microbial  communities  that  was  measured  in  soils  managed   by   phytoextraction   using   willow   short-­‐rotation   coppice   (SRC)   showed   no  significant  changes.  The  functional  diversity  of  the  soil  significantly  increased  only  in  soil  where  amendment  with  a  mixture  of  organic  matter  and  dolomitic  limestone  were  used  in   addition   to   the   willows,   indicating   that   microbial   communities   responded   to   SRC-­‐based   GRO   by   enriching   carbon   degradation,   nutrient   cycling   (nitrogen   and  phosphorous)  and  metal  resistance  response  gene  families.  

The   analysis   of   the  microbial   community   structure   by   the   clustering   of  DGGE  profiles  showed  clear  differences  between  the  treated  and  the  untreated  soils  (at  both  the  total  community  and  phylogenetic  group  level),  while  the  qPCR  technique  showed  differences  in   the   number   of   gene   copies   (nirK,   nirS,   nosZ,   amoA),   demonstrating   that   GRO  implementation   can   lead   to   a   shift   in   bacterial   community   and   diversity.   Shifts   in  community   structure   were   more   pronounced   in   soils   where   phytoexclusion   or  phytostabilisation  had  been  implemented.    Conclusions  Differences  between  the  treated  and  untreated  soils  were  detected  mainly  for  soils  that  were   treated   by   addition   of   soil   amendments,   especially   those   containing   organic  matter.   The   tests   that   showed   differences   include   chemical   extractions,   enzyme  activities,   soil   N   and   P   content,   microbiological   and   biochemical   endpoints,   the  functional  diversity  and  structure  of  the  soil  microbial  communities.    In   general,   the  microbiological   and   biochemical   endpoints   were   in   a   good   agreement  with  soil  toxicity  data,  TE  solubility  and  mobility  estimated  by  chemical  extractions,  and  can   be   robust   indicators   suitable   for  making   decisions   on  which  management   option  should  be  used.  Based  on  the  obtained  results  with  the  selected  evaluation  methods,  it  is  suggested  that  a  minimum  risk  assessment  battery  to  compare  or  monitor  the  sites  phytomanaged  by  GROs   might   consist   of   the   1M   NH4NO3   extraction   and   the   dwarf   bean   Plantox   test  including  the  plant  stress  enzyme  activities.    It  can  be  also  recommended  to  include  several  biochemical  analyses  into  test  batteries,  but  not  rely  on  a  single  endpoint,  especially  in  soils  where  organic  amendments  and  pH  conditioners   are   used.   This   is   because   soil  microbiological   and   biochemical   endpoints  respond   to   several   environmental   co-­‐variables   such   as   soil   organic   matter,   nutrient  content  and  availability,  pH  value,  water  holding  capacity,  plant  and  litter  cover.      Although   the   microbial   community   structure   showed   clear   differences   between   the  treated   and   the   untreated   soils,   these  methods   are   costly   and   labour   intensive   and   at  this  day  could  not  be  applied  on  a  commercial  basis.      3.4   Improving   GRO   through   plant   selection   and  modification   in   soil  trace  element  bioavailability      The   objectives   were   to   evaluate   and   identify   methods   or   tools   for   improving   the  performance  of  different  gentle   remediation  options.  The  work  carried  out  within   this  working   package   was   distributed   amongst   5   main   tasks,   targeting   the   following  strategies:   (1)   selection   of   appropriate   plant   species,   cultivars,   varieties   or   clones   for  application  in  distinct  GRO;  (2)  application  of  microbial   inoculants  for   improving  plant  performance  and/or  their  phytoremediation  potential;  (3)  addition  of  amendments  for  reducing  trace  element  (TE)  bioavailability  (in  situ  stabilisation  and  phytoexclusion,  and  (aided)   phytostabilisation);   (4)   incorporation   of   appropriate   agronomic   techniques   to  improve   plant   performance,   TE   removal   or   immobilisation;   and   finally,   (5)   the   use   of  effective  soil  amendments  and/or  amendment-­‐microbial  inoculant  combinations  for  TE  immobilisation  (phytostabilisation).  Selection   and   screening   of   plant   species,   cultivars,   varieties   or   clones   for  application  in  GRO  

For   phytoextraction   plants  must   be   able   to   accumulate   high   concentrations   of   TEs   in  their   harvested   parts   and   have   a   reasonably   high   biomass   production.   One   relevant  option   is   using   TE-­‐hyperaccumulators   which   are   able   to   accumulate   extreme  concentrations   of   metal(loid)s   in   their   above-­‐ground   biomass   and   at   the   same   time  possess  some  economic  added  value  (renewable  biomass   for  bio-­‐economy  and/or  bio-­‐ores).   High-­‐biomass   crops   (annuals   or   perennials)   and   woody   plants   (short   rotation  coppice   (SRC))   are   recognized   as   viable   alternatives   to   hyperaccumulators   for  phytoextraction  of  TEs  (particularly  Cd,  Se  and  Zn)  if  they  also  show  relevant  shoot  TE  removals   (i.e.   moderate-­‐high   bioconcentration   factor   (BCF)   and   high   shoot   yield).  Phytostabilisation   can   be   combined   with   excluder-­‐based   SRC   for   bioenergy   purposes  (selection  of  genotypes  is  based  on  TE  exclusion  and  also  based  on  their  characteristics  in   relation   to  conversion  processes,  e.g.   calorific  value,  bulk  density,  moisture  content,  ash   and   extractive   content).   Selection   of   pollutant-­‐excluding   agricultural   cultivars   for  cultivation  on  contaminated  and/or  remediated  land  contributes  towards  reducing  the  entrance  of  harmful  trace  elements  into  the  human  food  chain.    Various   plant   groups  were   assessed   throughout   the   duration   of   Greenland   on   both   a  field   scale   and   at   a   bench   level   (greenhouse   pot   experiments).   The   range   of   plant  types/species   assessed   by   the   different   groups   is   presented   in   Table   1.   Plant   species,  cultivars,   varieties   or   clones   were   assessed   for   distinct   GRO:   phytostabilisation   (TE  tolerance  and  exclusion  capacity),  in  situ  stabilisation  and  phytoexclusion  (TE-­‐excluding  phenotypes)   and   phytoextraction   (TE   accumulation   capacity).   Field/pot   experiments  were   established   and   carried   out   to   evaluate   plant   species   for   their   TE   resistance,   TE  extraction  potential  and/or  biomass  production.    In  particular,  well  assessed  mutant-­‐lines  of  sunflower  and  tobacco  with  enhanced  yield  and   stress   tolerance   were   provided   by   PT-­‐F   to   partners   for   further   assessment   and  development   at   the   different   Greenland   sites.   From   the   existing   seed   bank   of   230  sunflower   inbred-­‐lines   (M3-­‐M6),   the   most   promising   sunflower   mutant-­‐lines   were  selected   and   tobacco   in   vitro-­‐selections   were   performed.   Using   high   yielding   M3-­‐6  sunflower   mutant   inbred-­‐lines   with   large   genetic   variability   enables   more   efficient  mutant   selection   for   enhanced   TE   tolerance,   extraction   and   high   yield   characteristics,  and   speeds   up   fast   track-­‐breeding.   Four   successive   crops   of   sunflower   and   tobacco  variants  were  carried  out  in  five  Greenland  sites  over  the  period  2011-­‐2014.  The  plant  biomass   production   and   metal   extraction   potentials   for   the   different   variants   were  assessed   and   compared   between   sites   and   years.   The   progress   and   results   of   this  comparative  screening  were  presented  in  five  deliverables  (see  Table  3.4.1).    Table  3.4.1.  Plant  groups  assessed  by  different  partners  at  a  field-­‐  and  bench-­‐scale  throughout  Greenland.  

Plant   group  tested  

Partners  involved  

Related  deliverables  

Woody  plant  species:  Salix,   Populus  clones    

CSIC;  INRA;   UH;  LfULG  

D4.17:   List   of   Cu-­‐tolerant   grasses/trees   and   TE-­‐excluding  crop  cultivars  D4.21:  First  screening  of  Cu-­‐tolerant  grasses/trees  

Grassy  species:  Agrostis   sp.,    Festuca  sp.,    Vetiveria  

CSIC;  INRA;  IUNG;  

D4.17:   List   of   Cu-­‐tolerant   grasses/trees   and   TE-­‐excluding  crop  cultivars  D4.21:  First  screening  of  Cu-­‐tolerant  grasses/trees  

zizanioides,  Agropyron  elongatum    

LfULG  

High  biomass  annual  crops:  Tobacco,  sunflower  variants  

PT-­‐F;   UH;  INRA;  CSIC  

D4.16:   Provision   of   a   selection   of   improved   sunflower  and  tobacco  genotypes;    D4.19:  First  mutant  screening  on  pilot  sites;    D4.23:   Second   mutant   screening   and   candidate  selection;   D4.25:   First   comparative   mutant   screening  and   assessment;   D4.29:   Second   comparative   mutant  screening   and   assessment   of   top   ten   sunflower   and  tobacco  traits.    

Brassica  spp.   UH   -­‐  Agricultural  crops  (metal-­‐excluding  phenotype):  Barley,   wheat,  maize  cultivars  

AIT;  LfULG   D4.17:   List   of   Cu-­‐tolerant   grasses/trees   and   TE-­‐excluding  crop  cultivars  D4.20:   First   screening   of   TE-­‐excluding   agricultural  crops  

Hyperaccumulating  plant  species:  Noccaea  caerulescens  

CSIC;  BOKU  

-­‐  

 Agronomic   practices   for   improving   gentle   remediation   of   trace   element-­‐contaminated  soils  Upscaling   of   GRO   from   greenhouse   to   field   conditions   clearly   requires   incorporating  agronomical   knowledge   into   the   remediation   process.   The   influence   of   management  options,  such  as  planting  densities  and  harvest  systems,  or  the  need  for  crop  rotations  or  intercropping,   irrigation,   or   weed   and   pest   control,   are   not   taken   into   account   under  bench  scale  evaluations.  Based  on  the  experience  and  lessons  learnt  from  the  long-­‐term  Greenland  field  case  studies,  the  partners  involved  in  this  task  produced  a  review  article  summarising  agronomic  practices  against  their  demonstrated  or  potential  positive  effect  on   GRO   performance.   Potentially   negative   effects   of   GRO,   such   as   the   introduction   of  potentially   invasive   species,   were   also   discussed.   A   series   of   recommendations   were  made   for   increasing   the   GRO   success   and   aiding   stakeholders   in   related   decision-­‐making,  and  these  can  be  summarised  as  follows:    (1)   the   initial   spatial   variability   in   the   total   and   labile   TE   pools   should   be   well  characterized   in  order  to  enable  efficient   installation  and  monitoring  during   long-­‐term  trials.   The   same   is   true   for   the   distribution   of   the   (labile)   TE   pools   through   the   soil  profile   (and   this   is   pivotal   in   the   case   of   SRC).   Soil   (and   plant   samples)   should   be  archived  to  facilitate  any  retrospective  monitoring;    (2)  it  is  not  recommended  to  up-­‐scale  directly  from  studies  carried  out  at  a  bench-­‐level  (e.g.   pot   experiments)   to   large-­‐scale   site   applications.   At   least   one   step   on   an  intermediate   scale   should  be   conducted  on   site   (e.g.   several   small  plots  of   some  10   to  1000  m2)  and  if  possible  for  more  than  one  growth  season.  This  is  recommended  so  as  to   detect   any   potential   failures   due   to   long-­‐term   changes,   such   as   ageing   of   soil  amendments,   inter-­‐annual   changes   in   climatic   conditions,   pest   attacks,   litter   build-­‐up  and  release  of  dissolved  organic  matter,  changes  in  plant  and  animal  communities,  etc.).  

Additionally,   in  a  best  case  scenario   it  would  be  better   to  compare   in  parallel   the  best  GRO   and   conventional   technique   for   this   site   (for   better   demonstrating   the   pros   and  cons,  and  having  an  immediate  alternative  in  case  of  GRO  failure;    

(3)   soil   conditions   (e.g.   regarding   root   penetration,   water   retention,   organic   matter  content,  nutrient  supply,   factors  which  may   lead   to  plant   toxicity)  should  enable  plant  growth,  otherwise  intervention  (e.g.  sub-­‐soiling,  soil  amendments)  is  needed;    

(4)   weed   control   is   essential   during   the   early   establishment   of   plantations   due   to  competition  for  resources;    (5)   multi-­‐species/multi-­‐cultivar/multi-­‐clone   plantations   are   recommended   due   to  enhanced   plant   cover   resistance   against   unknown   or   unexpected   impacts   which  may  otherwise  lead  to  total  plant  loss  and  due  to  associated  benefits  related  to  pest  control  or   biodiversity.   The   biomass   production   should   be   considered   in   line  with   its   use   by  local  conversion  chains;    (6)  water  supply  vs.  water  requirements   is  vital  during  early  plant  establishment,  and  therefore  irrigation  should  be  considered;  (7)  fencing  or  some  means  of  protection  are  recommended  to  reduce  plant  loss  due  to  local  wildlife  herbivory,  several  clusters  being  better  than  only  one   large  fence  around  the  GRO  field  trial.      The  main  findings  related  to  this  task  formed  part  of  one  deliverable  (D4.22:  Knowledge  of   agronomic   practices   for   improving   plant   establishment)   and   were   published   in   the  joint  publication:  Petra   Kidd,   Michel   Mench,   Vanessa   Álvarez-­‐López,   Valérie   Bert,   Ioannis   Dimitriou,  Wolfgang   Friesl-­‐Hanl,   Rolf  Herzig,   Jolien  Olga   Janssen,  Aliaksandr  Kolbas,   Ingo  Müller,  Silke   Neu,   Giancarlo   Renella,   Ann   Ruttens,   Jaco   Vangronsveld,   Markus   Puschenreiter  (2015)   Agronomic   practices   for   improving   gentle   remediation   of   trace   element-­‐contaminated   soils.   International   Journal   of   Phytoremediation  (DOI:10.1080/15226514.2014.1003788,  in  press).    

Use   of   microbial   inoculants   for   improving   plant   performance   and/or  phytoremediation  potential  Exploiting   the  plant-­‐microbial  partnerships   in  phytoremediation   is   generally  based  on  the  capacity  of   the  bacteria   to,  on  one  hand,   improve  establishment,  growth  and  plant  survival   (plant-­‐growth   promotion);   and,   on   the   other   hand,   to   act   directly   on   the  contaminant.  Contaminated   sites  are  not  only  a   source  of   interesting  plant   species   for  application   in   phytoremediation   but   also   of   microorganisms.   During   GREENLAND  numerous  collections  of  plant-­‐associated  bacteria  were  obtained  from  a  series  of   trace  element-­‐contaminated   or   TE-­‐rich   soils   and   from   a   variety   of   plant   host   types   (from  hyperaccumulating  plant  species  to  high-­‐biomass  crops  and  woody  trees)  (Table  3.4.2).  Cultivable   strains   were   isolated   and   characterised   for   their   PGP   traits,   capacity   to  modify  soil  TE  bioavailability  and  their  potential  application  in  phytoremediation  of  TE-­‐contaminated  soils.    

 

Table   3.4.2.   Bacterial   collections   obtained   by   different   partners   during   the  duration  of  Greenland.  Based  on  D4.18:  List  of  microbial  strains  or  consortia  with  PGP  and/or  TE  solubilising/immobilizing  capacities.  Partner   Host  plant  species   Bacterial  type   Characterised  traits     References  

CSIC   Grass   species,   woody  trees  and  shrubs  

Rhizobacteria   16S   rDNA   sequence  identification,   Cd/Zn   MTC,  PO4,  Surf,  Sid,  IAA  

Becerra-­‐Castro   et  al.  2012  

CSIC   Hyperaccumulating  plant  species  

Rhizobacteria   16S   rDNA   sequence  identification,   Ni   MTC,   PO4,  Surf,  Sid,  IAA  

Álvarez-­‐López   et  al.  2015  

INRA   Agrostis  capillaris   Root  endophytes  

16S   rDNA   sequence  identification,  PO4,  Surf,  Sid,  IAA  

Kolbas  et  al.  2015  

UHASSELT  

Willow  clones   RhizobacteriaRoot  endophytes  

16S   rDNA   sequence  identification,   Cd  MTC,   PO4,  Surf,  Sid,  IAA,  OA  

Janssen  et  al.  2015  

UHASSELT  

Brassica  napus   Rhizobacteria  

Endophytes  

16S   rDNA   sequence  identification,   Cd  MTC,   PO4,  Surf,  Sid,  IAA,  OA  

Croes  et  al.  2015  

UHASSELT  

Agrostis  capillaris   Seed  endophytes  

16S   rDNA   sequence  identification,   Cd  MTC,   PO4,  Surf,  Sid,  IAA,  OA  

Truyens  et  al.  2014  

UHASSELT  

Lupinus  sp.   Rhizobacteria  

Endophytes  

16S   rDNA   sequence  identification,   Cd  MTC,   PO4,  Surf,  Sid,  IAA,  OA  

Weyens  et  al.  2014  

AIT   Zea  mays  cvs.   Rhizobacteria  

Endophytes  

16S   rDNA   sequence  identification,   Cd/Zn   MTC,  PO4,  Surf,  Sid,  IAA,  ACC  

Touceda-­‐González  et  al.  2015  

Abbreviations:  maximum  tolerable  concentration,  MTC;  PO4,   inorganic  PO4-­‐solubiliser;  Ac,  acid  producer;  Surf,  biosurfactant  producer;  Sid,  siderophore  producer;  IAA,  indoleacetic  acid  producer;  OA,  organic  acid-­‐producer  References:  C.  Becerra-­‐Castro,  C.  Monterroso,  A.  Prieto-­‐Fernández,  L.  Rodríguez-­‐Lamas,  M.  Loureiro-­‐Viñas,  M.J.  Acea,  P.S.   Kidd   (2012)   Pseudometallophytes   colonising   Pb/Zn   mine   tailings:   A   description   of   the   plant–microorganism–rhizosphere  soil  system  and  isolation  of  metal-­‐tolerant  bacteria.  J.  Hazard.  Mat.  217–  218:  350–  359.  Álvarez-­‐López,  V.,  Prieto-­‐Fernández,  A.,  Becerra-­‐Castro,  C.,  Monterroso,  C.,  Kidd,  P.S.  (2015)  Rhizobacteria  associated  with  the  flora  of  three  serpentine  outcrops  of  the  Iberian  Peninsula.  Plant  Soil  (submitted).  Kolbas,  A.,  Kidd,  P.,  Guinberteau,  J.,  Jaunatre,  R.,  Herzig,  R.,  Mench,  M.  (2015)  Endophytic  bacteria  take  the  challenge   to   improve   Cu   phytoextraction   by   sunflower.   Environ   Sci   Pollut   Res   (in   press),   DOI  10.1007/s11356-­‐014-­‐4006-­‐1.  Janssen,  J.,  Weyens,  N.,  Croes,  S.,  Beckers,  B.,  Meiresonne,  L.,  Van  Peteghem,  P.,  Carleer,  R.,  Vangronsveld,  J.  (2015)  Phytoremediation  of  metal  contaminated  soil  using  willow:  exploiting  plant-­‐associated  bacteria  to  improve  biomass  production  and  metal  uptake.  International  Journal  of  Phytoremediation  (submitted).  Truyens,  S.,  Jambon,  I.,  Croes,  S.,  Janssen,  J.,  Weyens,  N.,  Mench,  M.,  Carleer,  R.,  Cuypers,  A.,  Vangronsveld,  J.  (2014)   The   effect   of   long-­‐term   Cd   and  Ni   exposure   on   seed   endophytes   of  Agrostis   capillaris   and   their  potential   application   in   phytoremediation   of   metal-­‐contaminated   soils.   International   Journal   of  Phytoremediation,  16,  643–659.  DOI:  10.1080/15226514.2013.837027  Weyens,   N.,   Gielen,  M.,   Beckers,   B.,   Boulet,   J.,   van   der   Lelie,   D.,   Taghavi,   S.,   Carleer,   R.,   Vangronsveld,   J.  (2014)  Bacteria  associated  with  yellow  lupine  grown  on  a  metal-­‐contaminated  soil:  in  vitro  screening  and  in   vivo   evaluation   for   their   potential   to   enhance   Cd   phytoextraction.   Plant   Biology,   16:   988-­‐996.  doi:10.1111/plb.12141  Croes   S.,   Weyens   N.,   Colpaert   J.,   Vangronsveld   J.   (2015)   Characterization   of   the   cultivable   bacterial  populations   associated   with   field   grown   Brassica   napus   L.:   an   evaluation   of   sampling   and   isolation  protocols.  Environmental  Microbiology,  doi:10.1111/1462-­‐2920.12701.  Touceda-­‐González,  M.,   Brader,   G.,   Antonielli,   L.,   Balakrishnan  Ravindran,  V.,  Waldner,   G.,   Friesl-­‐Hanl,  W.  Sessitsch,  A.  (2015)  Bioavailability  of  heavy  metals,  plant  growth  and  microbiome  characteristics  due  to  the  amendment  of   immobilizers  and  the  plant  growth-­‐promoting  strain  Burkholderia  phytofirmans  PsJN.  Soil  Biology  and  Biochemistry  (submitted).    

A   large   part   of   the   Greenland   project   was   dedicated   to   carrying   out   bench-­‐scale  bioaugmentation   trials   using   different   combinations   of   potential   PGP/metal-­‐(im)mobilising   bacterial   strains   and   plant   species.   Promising   plant-­‐bacterial  partnerships   were   presented   in   the   deliverable   D4.24:   Testing   of   candidate   plant-­‐microbial  partnerships.  In  general  these  studies  showed  that  the  inoculation  of  selected  plant  species/cultivars/clones  with  beneficial  plant-­‐associated  bacteria  can  significantly  enhance  biomass  production  and/or  plant  metal  uptake,  but  that  the  results  are  highly  plant-­‐,  bacterial  strain-­‐  and  soil-­‐specific.  Some  advances  were  also  made  in  small-­‐scale  field  trials,  these  studies  underlined  the  need  for  further  research  and  also  showed  that  results   obtained   in   in   vitro   or   using   pot-­‐scale   experiments   do   not   always   correspond  with  the  effects  obtained  in  the  field  where  additional  complicating  factors  can  influence  the  overall  outcome  and  performance  of   the   inoculant.  The  main   results  obtained  at  a  field-­‐scale   were   presented   in   deliverable   D4.26:   Identification   of   microbial   strains   or  consortia  with  PGP  and/or  TE  solubilising/immobilising  capacities  at  field-­‐scale.  

In  addition  to  the  publications  of  the  individual  partners  (listed  in  previous  reports  and  D6.40)  some  joint  publications  resulted  from  this  task:  

Angela   Sessitsch,   Melanie   Kuffner,   Petra   Kidd,   Jaco   Vangronsveld,   Walter   W.   Wenzel,  Katharina  Fallmann,  Markus  Puschenreiter  (2013)  The  role  of  plant-­‐associated  bacteria  in   the   mobilization   and   phytoextraction   of   trace   elements   in   contaminated   soils.   Soil  Biology  &  Biochemistry  60:  182-­‐194.  

Álvarez-­‐López,  V.,  Prieto-­‐Fernández,  A.,  Janssen,  J.,  Herzig,  R.,  Vangronsveld,  J.,  Kidd,  P.S.  (2015)   Bacterial   inoculation   methods   influence   the   phytoextraction   capacity   of  Nicotiana  tabacum.  International  Journal  of  Phytoremediation  (in  preparation).  

Use   of   amendments   for   reducing   TE   bioavailability   (in   situ   stabilisation   and  phytostabilisation)  

Soil   amendments   including   liming   agents   (calcite,   burnt   lime,   slaked   lime,   dolomitic  limestone),   phosphates   and   apatites,   Fe,   Al   and  Mn   oxyhydroxides,   biochars   (carbon-­‐rich  end  product  of  the  pyrolysis  of  biomass),  organic  amendments,  and  industrial  waste  products  have  been  widely  used  in  phytostabilisation  and  some  (aided)  phytoextraction  experiments.  The  formation  of   insoluble  TE  chemical  species  reduces  leaching  through  the  soil  profile  and  the  labile  metal  pool  in  the  soil.  Amendment  trials  were  carried  out  throughout  the  full  duration  of  the  Greenland  project  by  the  different  partners  on  both  a  field-­‐   and   bench-­‐scale   (see   Table   3.4.3).   In   addition,   several   collaborative   initiatives  were   carried   out   between   WP4   members,   and   one   collaborative   initiative   between  Greenland   and   the   FP7   HOMBRE   project   (Nº.   265097).   The   main   results   of   these  screenings  were   presented   in   the   deliverable  D4.27:  Database  of  efficient  amendments  and/or   combinations   of   amendments   for   use   in   phytostabilisation   approaches   and  selection  of  treatments  for  particular  site  conditions.  The  effectivity  of  soil  amendments  is  generally  assessed  on   the  basis  of  physico-­‐chemical  and  selective  chemical  extractions  demonstrating   a   reduction   in   soil  metal  mobility   and   availability.   However,   biological  evaluations   are   also   vital   when   assessing   the   potential   use   of   a   soil   amendment   in   a  given   remediation   procedure.   In   these   studies   biochemical   properties   (soil   enzyme  activities)  as  well  as  an  ecotoxoxicity  assay  (using  earthworms)  was  carried  out  in  order  to   obtain   a   more   reliable   estimation   of   the   efficiency   of   the   metal   immobilisation  obtained.   Effective   amendments   were   identified   for   the   in   situ   stabilisation   and  phytostabilisation   of   Cd-­‐,   Pb-­‐   and   Zn-­‐contaminated   soils.   Results   indicated   that  amendment  combinations  were  the  most  successful  in  reducing  toxicity  and  promoting  plant   growth   and   soil   enzyme   activities,   and   reducing   metal   mobility   and  

bioaccumulation   in   earthworms.   The   best-­‐perfomring   combinations   for   this   type   of  contaminated   soil   were   the   CaHPO4+drinking   water   residues   (DWR)+compost,   iron  grit(IG)+Linz-­‐Donawitz  slag+compost,  and  IG+cyclonic  ashes+compost.    Table   3.4.3   Soil   amendments   used   for   in   situ   stabilisation   and   (aided)  phytostabilisation  and  assessed  during  the  Greenland  project  

Inorganic  amendments   Organic  amendments  Rock   phosphate   (a   major   source   of   P  fertilizers)  

Manures    

Thomas   basic   slag   (a   by-­‐product   of   the  iron  industries)  

Biosolids   (sewage   sludge),   Composted  biosolids    

Wood  ashes   Green  waste  composts  Cyclonic  ashes    Zerovalent  iron  grit   Others  Linz-­‐Donawitz  slag   Biochar  Siderite    Gravel  sludge    Red  mud    Drinking  water  residues      Bench-­‐scale   evaluations  were   carried   out   to   assess   the   potential   of   different   biochars  and  green  waste  compost  to  immobilise  Cu  in  a  contaminated  soil.  Biochar  and  compost  were  shown  to  reduce  leachable  copper  in  contaminated  soil,  the  combined  application  of  the  two  amendments  was  shown  to  be  effective  for  immobilisation  and  plant  growth,  biochars   and   composts  were   shown   to   increasingly   immobilise   copper  with   increased  application  rate,  and  the  amendments  do  not  effect  a  reduction  in  copper  bioavailability  alone,   but   rather   initiate   multiple   concomitant   changes   to   soil   which   contribute   to  reduced   phytotoxicity.   These   results   indicate   that   these   amendment   types   could   be  successfully  used   in   combination  with  phytoremediation   to   further  decrease  pollution  risks   and   potentially   provide   a   saleable   energy   crop.   However,   more   research   is  required  to  further  establish  the  detailed  operating  windows  of  these  amendments  and  to   more   clearly   define   the   influence   of   different   feedstock   materials   on   biochar   and  recycled  organic  matter  properties.  Additionally,   the   amendments   successfully   trialled  will   require   field   trials   to   determine   their   efficacy   on   a   larger   scale   and   confirm   their  potential   for   deployment   on   a   full-­‐scale   remediation   site.   Nonetheless,   both   recycled  organic   matter   and   biochar   have   the   potential   to   simultaneously   stabilise   soil  contaminants,  improve  soil  quality  and  offer  carbon  sequestration  benefits.  Biochar  has  been  shown  to  have  significant  longevity  and  therefore  may  be  economically  attractive,  as  it  may  provide  a  long-­‐term  effect  without  repeat  applications.  Finally,   GREENLAND   assessed   the   combined   use   of   TE-­‐immobilising   soil   amendments  and  beneficial  plant-­‐associated  bacteria  as  a  means  of   simultaneously   improving  plant  growth/establishment   and   reducing  metal  mobility   and  bioavailability.  Again   this  was  assessed  on  a  bench-­‐  and  field-­‐scale,  and  presented  in  the  deliverable  D4.28:  Report  on  possibility  to  enhance  assisted  phytostabilisation  through  microbial  inoculation.    3.5  Appraisal  of  current  GRO  practice,  and  development  of  implementation  guidance  and  decision  support  

This  document  summarises  the  main  S&T  results  /  foregrounds  delivered  by  WP5  of  the  GREENLAND   project   (Appraisal   of   current   GRO   practice,   and   development   of  implementation  guidance  and  decision  support).  Results  are  presented  in  relation  to  WP  deliverables.        D5.1:  (Multi-­‐lingual)  Best-­‐practice  guidance  document  for  the  application  of  GRO  at   field-­‐scale   (including  appraisal   of   the   various  options   available,   evaluation  of  large-­‐scale   field   trials,   analysis  of  valorisation  potential,   and  suggested  methods  and  monitoring)    Outputs  from  WP1-­‐4  were  collated,  evaluated  and  reviewed  to  produce  a  best-­‐practice  guidance   document   for   the   application   of   GRO   at   field-­‐scale.   Following   detailed  discussions   at   the   regular   Greenland   project   meetings   (including   with   the   project  Advisory  Board  members),  and  critical  review  of  existing  guidance  documents  from  (a)  project   partners   AIT   and   BOKU,   and   (b)   ADEME   (Fr),   a   first   draft   of   the   guidance  document   was   tabled   at   the   October   2013   periodic   meeting   (Brighton).   A   finalised  version   of   the   Best   Practice   Guidance   document   was   completed   in   December   2014,  following  tabling  (and  input  from  Greenland  partners  and  Advisory  Board  members)  at  the  Frankfurt  end  of  project  meeting.  Versions  of  the  document  have  been  produced  in  English,  French  and  German,  with  the  Best  Practice  Guidance  now  available  through  the  Greenland  project  website  (www.greenland-­‐project.eu)  and  via  the  EUGRIS  web  portal  (www.eugris.info).     The   Best   Practice   Guidance   document,   which   cross-­‐links   to   the  project  decision  support  tool  (DST,  deliverable  5.3),  is  structured  as  follows:      

1. Definitions  and  context  –  what  is  GRO  and  how  does  it  work?  2. Overview  of  current  state  of  development  and  risk  management  capability  3. Case  /  success  stories  4. Potential  economic,  environmental  and  social  benefits  5. Operating  windows  for  GRO  6. Further  information  sources  

Appendices:    Appendix  1:  Design  and  implementation  (WP1)  Appendix  2:  Cultivars  and  amendments  (WP4)  Appendix  3:  Safe  biomass  usage  (WP2)  Appendix  4:  Indicators  of  success  and  methods  (WP3)  Appendix  5:  DST  and  cost-­‐calculator  (WP5)  Appendix  6:  Stakeholder  engagement  guidelines  (WP5)  Appendix  7:  Further  examples  and  case  studies  (WP1)    Within   the   document,   section   3   presents   three   case   studies   from  WP1,   representing  cases   where   application   of   phytoextraction,   aided   phytostabilisation,   and   in   situ  stabilisation   /   phytoexclusion   phytomanagement   strategies   have   led   to   demonstrable  source   removal,   pathway   management   or   receptor   protection.   Further   examples   are  given   in  Appendix  7.   In   section  4,   the   guidance   cross-­‐refers   to   three   assessment   tools  which   allow   assessment   of  wider   benefits   from  GRO   application:   the   European  Union  FP7  HOMBRE  project  (grant  265097,  www.zerobrownfields.eu)  Brownfield  Opportunity  Matrix  (BOM);  the  SURF  indicator  sets  on  sustainability;  and  an  outline  Cost-­‐Calculator  (developed   within   the   Greenland   project).   These   tools   were   felt   to   provide   a   more  thorough   and   defensible   assessment   of   wider   benefits   from   GRO   application   than   a  previous   tabular   indicator   set   approach   used   in   earlier   draft   versions   of   the   Best  

Practice   Guidance,   and   are   discussed   further   under   deliverable   5.3.   Under   section   5  (Operating   Windows)   of   the   guidance   we   present   quick   reference   tables   on   GRO  applicability   (Are   GRO   applicable   to   your   site?)   and   treatable   contaminants   (Which  metal(loid)   contaminants   can   GRO   treat?),   and   a   link   to   outline   Operating   Window  Matrices  in  the  project  DST  (see  Deliverable  5.3).            D5.2:  Guidelines  for  stakeholder  participation,  engagement  and  empowerment  when  implementing  GRO    Following  collation  and  review  of  existing  methods  of  stakeholder  engagement  (MS34,  submitted  at  end  of  project  year  1),  and  stakeholder  engagement  discussions  held  at  the  periodic   Greenland   meeting   in   Pulawy   and   at   the   9th   International   Conference   on  Phytotechnologies   at  Hasselt   University   in   September   2012   (via   a   special   round-­‐table  session),   a   review   publication   with   recommendations   and   stakeholder   engagement  guidelines   for  GRO  was   published   in   the   Journal   of   Environmental  Management   (JEM,  Impact  Factor  3.245)  in  2013.    This  review  article  (co-­‐authored  by  representatives  of  6  Greenland  partner  organisations)  incorporates  case  study  material  from  the  Greenland  test  sites  Biogeco  platform  (SW  France),  and  Krummenhennersdorf  (Saxony,  Germany),  and   reviews   the   current   state   of   the   art   in   stakeholder   engagement   strategies.   It  identifies   key   principles   for   stakeholder   engagement   when   implementing   GRO.  Following  publication  of  this  article,  work  under  this  deliverable  focused  on:      

1. Integrating   stakeholder   engagement   strategies   and   guidance   into   the   project  decision  support  tool  (DST)  and  the  Best  Practice  Guidance;    

2. Developing   criteria   for   the   identification   of   different   stakeholders  profiles/categories   -­‐   their   expectations,   influence,   characteristics,   preferred  approaches  to  engagement  and  levels  of  engagement.  

 Under  these  points,  a  stakeholder  engagement  module  has  been  included  in  phase  2  of  the  project  DST  which  gives  a  context  and  rationale  for  stakeholder  engagement,  shows  (a)  methods   and   strategies   for   effective   stakeholder   engagement   and   (b)   stakeholder  classification  criteria,  and  which  presents  the  key  guidance  for  stakeholder  engagement  published  in  the  Journal  of  Environmental  Management  paper.  Under  point  2  specifically  a   table   and   representative   listing   for   identification   and   classification   of   stakeholders,  based  on  published  literature  and  current  practice  at  the  Greenland  and  other  sites,  has  been  presented  and   finalised  at   the   final  project  meeting   in  Frankfurt,  and   included   in  the  stakeholder  engagement  module  of  the  DST.  Key  information  from  the  stakeholder  engagement   module,   including   stakeholder   engagement   guidelines   for   GRO,   is   also  presented  in  Appendix  5  of  the  Best  Practice  guidance.            D5.3:  Practical  GRO-­‐focused  decision  support  tool  (DST),  for  integration  into  existing  national  decision  support  frameworks    A   working   GRO-­‐focused   DST   was   developed   in   the   early   project   phases   (deliverable  MS35  –  Production  of  Outline  (Generic)  DST),  following  the  “tiered”  (or  layered)  model  proposed   in   the   earlier   ERA-­‐NET   SUMATECS   project   and   in   Onwubuya   et   al.   (2009),  initially  within  the  framework  of  the  UK  Model  Procedures  for  the  Management  of  Land  Contamination   (CLR11).   Research   underpinning   this  DST  development  was   submitted  as   a   PhD   thesis   by   K.   Onwubuya   (University   of   Brighton)   in   early   2013,   which   also  includes  validation  and  testing  of   this  outline  model  using  3  case  studies  (sites   in  east  

London,  Lommel,  and  the  Biogeco  platform),   the   latter   two  drawn  from  the  Greenland  test  sites.  This  thesis  was  successfully  defended  in  April  2013.  Following  this  testing  and  evaluation  of  the  outline  model  produced  (milestone  MS36)  the  project  DST  was  further  developed   and   populated   during   year   4   of   the   project,   drawing   on   outputs   from   the  other  Greenland  WPs  (1-­‐4),  and  following  tabling  (and  feedback)  at:        

(a) the  11th   International   Conference  of   the   International   Phytotechnology   Society  (Sept  30th  –  Oct.  3rd  2014,  Heraklion,  Crete),  to  a  dominantly  European  and  US  academic  and  industry  audience;  and    

(b) the   “CABERNET   2014:   Tailored   &   Sustainable   Redevelopment   -­‐   towards   Zero  Brownfields”   conference   (14-­‐16th   October   2014   in   Frankfurt   am   Main,  Germany),   to   a   dominantly   EU-­‐based   research,   industry   and   consultancy  audience,  

and   following   feedback   from   project   Advisory   Board  members   at   the   final   Greenland  project  meeting  (also  Frankfurt).      The  DST  is  a  3  phase  MS  Excel-­‐based  model  (figure  1),  designed  to  build  in  complexity  and  time  effort  (and  technical  detail)  through  its  3  phases,  and  which  embeds  the  best  practice   guidance   produced   in   D5.1   and   the   stakeholder   engagement   principles   and  identification   criteria   produced   in   D5.2.   It   is   designed   to   link   to   existing   national  decision  support  frameworks  at  the  Options  Appraisal  stage.  A  full  user's  guide  for  the  tool  can  be  accessed  by  selecting  a  "User  Guide"  tab  on  the  front/entry  page  (figure  1).  Phase   1   (Initial   Concepts/Feasibility)   includes   definitions,   high   level   operating  windows   for   GRO   (in   terms   of   their   scope   and   risk  management   capability),   “success  stories”   of   GRO   (i.e.   where   GRO   strategies   have   led   to   demonstrable   source   removal,  pathway  management   or   receptor   protection),   and   an   outline   contaminant  matrix   for  the   applicability   of   various   GRO   options   at   trace   element   contaminated   sites   (TECS),  based   on   data   from   the   GREENLAND   site   network.   Phase   2   (Exploratory  stages/confirmation)   includes   modules   on   stakeholder   engagement   (including  stakeholder   engagement   principles   and   stakeholder   identification   criteria,   discussed  under  D5.2  above),  and  sustainability/  wider  benefits  assessment,  where  links  to  three  matrices/modules  are  provided:  (a)   The   European   Union   FP7   HOMBRE   project   (grant   265097,  www.zerobrownfields.eu)   Brownfield   Opportunity   Matrix   (BOM).   This   is   an  Excel-­‐based   qualitative   screening   tool   to   help   decision   makers   identify   which  services   they   can  obtain   from   “soft   reuse”   interventions   (including  GRO)   at   a   site,  and   how   these   services   interact.   Greenland   partners   have   collaborated   with  HOMBRE   to   populate   the   operating   and   opportunity  windows   for  GRO  within   the  BOM.   The  matrix   can   be   used   to  map   the   prospective   range   of   opportunities   that  might   be   realised   by   a   remediation   or   redevelopment/regeneration   project   for  “soft”  reuses1,  and  the  project’s  consequent  sources  of  value.    

(b)   The   SURF   indicator   sets   on   sustainability,   which   outline   the   various   headline  indicator  categories   that  should  be  considered  during  sustainability  assessment   in  land   remediation   projects.   These   indicator   categories   provide   a   checklist   for  agreeing  a  scope  for  a  sustainability  assessment.    A  tiered  approach  to  sustainability  assessment   is   suggested,   in  keeping  with  guidance   from  NICOLE  (www.nicole.org)  and   SuRF-­‐UK.     Many   decisions   may   be   resolvable   on   the   basis   of   a   qualitative  

                                                                                                               1  I.e.  where  the  soil  surface  is  not  sealed  by  a  building  or  infrastructure  

approach.    Where  issues  are  more  complex  a  semiquantitative  or  even  quantitative  approach   may   be   used   such   as   Life   Cycle   Assessment   (LCA)   and   Cost   Benefit   or  Multi-­‐Criteria  Analysis  (CBA/MCA).    However,  in  most  caes  quantitative  approaches  only  allow  a  partial  consideration  of  the  full  range  of  possible  sustainability  issues.  

(c)   An   outline   Cost-­‐Calculator,   which   has   been   developed   within   the   Greenland  project   and   incorporates   user-­‐entered   cost   data   (including   site   preparation   costs;  plant   and   planting   costs;   site   costs;   biomass   costs   and   revenues;   and  monitoring  costs)   to  estimate   the  economic  value  proposition  of  GRO  at  a  particular  site.  This  module  has  been  “calibrated”  using  data  from  the  Greenland  site  network,  which  are  used  to  test  the  cost  calculator  and  give  input  examples.  

 Phase   3   (Design   Stages)   of   the   DST   provides   a   technical   assessment   and  implementation  guidance  for  GRO,  with  detail  from  WP1  –  4,  and  also  outline  operating  windows   for  GRO.   For   the   latter,  we  provide   three  MS  Excel-­‐based  operating  window  matrices   which   allow   the   user   to   check   the   outline   applicability   of   GRO   (grouped   as  phytoextraction,   phytostabilisation,   and   immobilisation/phytoexclusion)   to   a   specific  site,   in   terms   of   local   soil   pH,   site   plant   toxicity,   climate,   soil   type,   and   depth   of  contamination.   The   tool   then   refers   the   user   to   national   contact   points   from   the  Greenland  consortium,  and  further  references.      

 

   Figure  1  :  Front/entry  page  of  the  Greenland  project  decision  support  tool  (DST),  showing  overall  DST  structure  and  introductory  text  for  the  user.    The   project   DST   has   been   finalised   and   uploaded   to   the   Greenland   project   website  (www.greenland-­‐project.eu),   and   forms   the   focus  of   a   technical  paper  which  has  been  submitted   to   the   journal   “Remediation”,   targeting   this   journal’s   main   audience   of  contaminated  land  regulators,  consultants  and  practitioners.      

 4.  The  potential  impact  (including  the  socio-­‐economic  impact  and  the  wider  societal  implications  of  the  project  so  far)  and  the  main  dissemination  activities  and  exploitation  of  results      Soil  is  a  non-­‐renewable  natural  resource.  It  provides  the  basis  for  all  terrestrial  human  activities,   including  agriculture  and  forestry,  and  is  therefore  an  irreplaceable  basis  for  the  survival  and  prosperity  of  the  European  society.  Due  to  anthropogenic  activities  (e.g.  industrialisation,   intensive  agriculture,  urbanisation  etc.)   soil   is   getting   lost   every  day.  Only  by   soil   sealing,  275  hectares  of   soil  were   lost  per  day   in   the  1990s,  with  a   slight  reduction  to  252  hectares  per  day  in  recent  years.  In  addition  to  the  loss  of  soil,  further  soil   resources   are   degraded   by   organic   matter   decline,   salinization   or   pollution.   The  Thematic  Strategy  for  soil  protection   lists  soil  contamination  as  one  of   the  eight  major  threats   to   soil.   Local   soil   pollution   has   caused   localised   hot   spots   of   contamination,  whereas  diffuse  soil  pollution  has   lead  to  widespread  contamination  of  soils  on  a  very  large  scale.      Conventional   remediation   technologies   can   (at   very   high   costs)   decontaminate   small  scale  contaminations,  whereas  moderately  polluted  soils  on  a  much   larger  scale   (km2)  remained   mostly   untreated   because   of   the   impossibility   to   treat   millions   of   m3   of  contaminated  soils  with  excavation  or  washing  treatments.  GRO  may  serve  as  low  cost,  environmental   and   socio-­‐economic   friendly   solutions   to   remediate   polluted   soils   on   a  very   large   scale   and   thus   help   to   overcome   previous   limitations.   In   the   GREENLAND  project   the   remaining   problems   of   GRO   development,   up-­‐scaling   and   biotechnological  improvement   have   been   successfully   addressed   and   a   set   of   well-­‐proven   methods  allowing  the  remediation  of  TE-­‐contaminated  soil  at   low  cost  and  with  the  potential  to  deliver   additional   financial   benefits   for   land   owners   were   delivered.   In   this   way,  GREENLAND  delivered  GROs  as  a  basis   for  a  new  generation  of  green  soil  remediation  technologies   and   thereby   significantly   contributed   to   overcoming   one   of   the   most  problematic   threats   to   European   soils.   More   widely,   promoting   sustainable   soil  remediation  methods,  such  as  GRO,  contributes  to  a  number  of  the  priorities  listed  in  the  Lisbon   and   Gothenburg   agendas,   such   as   climate   change,   public   health   and   resource  management   as   well   as   enhancing   research   and   technological   development   and  promoting  entrepreneurship  and  skills.    Contamination   of   soils   with   TE   is   still   a   major   problem   in   Europe.   According   to   the  European   Environmental   Agency,   up   to   3   million   sites   in   Europe   are   contaminated,  approximately  37  %  of   them  with  TE.  This  European  problem  can  only  be  solved  on  a  European   scale.   The   best   practice   guidance   document   produced   by   GREENLAND  (http://www.greenland-­‐project.eu/downloads/Greenland_Handout.pdf)   provides   all  the  necessary   information  for  applying  GRO  successfully  and  the  decision  support   tool  (DST;   http://www.greenland-­‐project.eu/downloads/DST%20download.xlsx)   allows  selecting   the   most   suitable   GRO   for   the   specific   site   conditions   and   offers   further  information   on   stakeholder   engagement   and   empowerment   for   further   enhancing   the  success  of   the  GRO  application.  The  best  practice  guidance   is  also  available   in  German  (http://www.greenland-­‐project.eu/downloads/Greenland_best-­‐practice_guide_German.pdf)   and   French   (http://www.greenland-­‐project.eu/downloads/Greenland_best-­‐practice_guide_French.pdf).     Considering   the  huge  number  of   contaminated   sites,  GRO   implementation  offers   also   a  huge   economic  

potential   for   new   entrepreneurship   in   bio-­‐economy.   New   companies   can   be   founded,  focusing   on   GRO   as   a   new   and   innovative   environmental   technology,   applied   on  contaminated   arable   land   (e.g.   in   cooperation   with   local   farmers   and   farmer’s  associations),  mine  tailings,  but  also  on  former  industrial  sites,  i.e.  so-­‐called  brownfields.  In   the   latter   case,   GRO   application  might   be   integrated   into   a   general   brownfield   re-­‐development  plan  (further  information  is  provided  by  the  Brownfields  Management  and  Avoidance   brochure,   developed   and   published   by   the   HOMBRE   FP7   project:  http://www.zerobrownfields.eu/quicklinks/HOMBRE_Broschure_2014_FINAL.pdf).  GRO   application   includes   in   most   cases   the   use   of   specific   plants   (and   associated  microbes).   In   this   way,   biomass   is   produced,   offering   the   opportunity   of   using   it   in  various  ways.  However,  in  some  cases  the  biomass  is  enriched  with  trace  elements,  thus  GREENLAND   focused   on   biomass   valorisation   options   that   are   suitable   for   biomass  harvested  on  GRO  sites.  By  applying  the  tested  technologies  (e.g.  combustion,  pyrolysis,  etc.)   energy   and   raw   materials   for   further   use   are   produced.   Due   to   the   increased  demand  for  biomass  in  Europe,  e.g.  because  of  the  substitution  of  fossil  fuels  or  the  need  for   raw  materials   produced   from   plant   biomass,  more   and  more   sites   are   needed   for  non–food   biomass   production.   In   order   to   avoid   a   competition   for   land   between  production  of  non-­‐food  biomass  and  food/fodder,  additional  sites  are  needed.  Marginal  lands,  e.g.  TE-­‐contaminated  sites,  could  offer  a  great  opportunity  for  providing  sites  for  biomass  production,  while  due  to  the  application  of  GRO  the  sites  are  remediated.      The  output  of  GREENLAND  has  been  presented  to  different  target  groups  in  many  ways.  The   general   public   has   been   addressed   in   several   newspaper   articles   as  well   as   radio  and  TV  broadcasts.  The  GRO  performance  in  the  field  has  been  demonstrated  in  several  field  days,  where  the  local  community  (e.g.  farmers,  local  decision  makers,  local  media)  had   the   opportunity   to   see   GRO   sites   and   to   talk   with   site   managers.   On   one   site  (Arnoldstein,  Austria)  also  school  classes  (Fig.  4.1)  were  invited  to  experience  how  GRO  works  in  practice.  The  role  of  soils  as  a  natural  resource,  loss  of  soil  and  remediation  of  soils   were   discussed   at   the   site,   but   also   in   associated   workshops   in   their   schools.  Newspaper  articles  as  well  as  radio  and  TV  reports  provided  information  on  GRO  for  the  local   population.   On   the   GREENLAND   website   (www.greenland-­‐project.eu)   general  information  and  project  results  (e.g.  case  studies,  success  stories,  best  practice  guidance  document,  DST)  were  offered  for  scientists,  stakeholders,  practitioners  and  the  general  public.          

   

     Figure   4.1:   Left:   School   children   visiting   the   site   Arnoldstein;   Austria   (source:  http://www.nms-­‐noetsch.ksn.at/).  Right:  Field  day  at   the  site  Lommel,  Belgium,  at   the  harvesting  (source:  www.niewsblad.be,  25.2.2014).    One  major  target  group  of  GREENLAND  included  stakeholders  and  decision  makers  on  a  national   level.   They   were   addressed   in   two   major   ways:   on   the   one   hand,   six  representatives   of   national   environmental  ministries   or   environmental   agencies  were  invited   to   join   the   advisory   board   (AB)   of   GREENLAND.   This   board   consisted   of  representatives  of  the  following  organisations:    

• AT:  Austrian  Environmental  Agency  (Umweltbundesamt);  Gernot  Döberl  • DE:  German  Enviromental  Agency  (Umweltbundesamt):  Jörg  Frauenstein  • FR:   Agence   de   l’Environnement   et   de   la   Maîtrise   de   l’Énergie   (ADEME):  

Frédérique  Cadiere  • SE:   2011-­‐2012:   NATURVÅRDSVERKET:   Inger   Johansson;   2013-­‐2014:    

Länsstyrelsernas  tillsynssamordnare  förorenade  områden:  Klas  Köhler  • PL:  Polish  Ministry  of  Environment,  Department  of  Environmental   Instruments:  

Joanna  Kwapisz    

They   took  part   in   four  project  meetings,  where   the  project   output  was  presented   and  discussed.   The   AB   provided   important   feedback   regarding   the   project   progress   and  output,   especially   in   the   context   of   the   different   national   circumstances.   Furthermore,  the   AB   offered   substantial   support   for   the   development   of   the   DST   and   the   practical  guidance  handbook.  In  particular  the  handbook,  showing  not  only  technical  instructions,  but  also  success  stories  for  GRO  application,  was  considered  by  the  AB  as  a  crucial  tool  for  bringing  GRO  into  practical  application.  In  addition,  they  also  acted  as  “information  hubs”,   transmitting   the   Greenland   information   to   other   stakeholders   and   decision  makers  (e.g.  the  Swedish  EPA).    On   the   other   hand,   national   and   regional   stakeholders   as   well   as   consultancies   and  environmental   enterprises   were   also   addressed   in   several   local   workshops   and  meetings,  e.g.,  R&D  days  on  management  of  polluted  sites  and  soils  Paris  2012  and  Paris  2014;   INRA   Ecotoxicology   Network   –   Biarritz   2014;   several   meetings   with   a  representative  (Dr.  Harald  Kasamas)  of  the  Austrian  Ministry  for  Environment  and  the  Austrian   Society   for   Management   of   Landfills   and   Contaminated   Sites   as   well   as   the  Austrian  Soil  Science  Society;  and  many  more.    

 The   results   of  Greenland  were   also  presented   at   several   scientific   conferences,  with   a  total   number   of   250   oral   and   poster   presentations   (see   list   of   dissemination).   In   this  context,  one  major  event  was  the  joint  final  conference  of  the  FP7  projects  GREENLAND,  HOMBRE,   TIMBRE   and   GLOCOM   in   Frankfurt   am   Main   (Germany)   on   October   14-­‐17  2014   (http://www.zerobrownfields.eu/Displaynews.aspx?ID=566).   In   this   forum,   the  main  output  of  GREENLAND  was  presented  to  a  diverse  audience  of  scientists,  students  and   stakeholders.   Together  with   the   results   of   the   other  projects,   all  wider   aspects   of  soil   remediation,   brownfield   regeneration   and   associated   topics   such   as   biomass   use  and  valorisation  were  presented  and  discussed.  In  addition,  the  results  of  GREENLAND  were  prominently  presented   in  several   relevant   international  conferences,   such  as   the  International   Phytotechnology   Conferences,   the   SETAC   (Society   of   Environmental  Toxicology   and   Chemistry)   conferences,   the   International   Conferences   of   the  Biogeochemistry  of  Trace  Elements  (ICOBTE),  the  Sustainable  Remediation  Conference,  and  many  more.  Within  these  conference  series,  GREENLAND  has  become  a  well-­‐known  “trademark”  for  research  work  on  GRO.    Finally,  the  output  of  GREENLAND  was  published  in  several  scientific  journals.  In  total,  114   papers  were   published   (see   dissemination   list).   The  most   important   publications  were  focussing  on  the  assessment  of  GRO  efficiency,  on  agronomic  measures  to  improve  GRO  efficiency  and  on  the  decision  support  tool  and  stakeholder  engagement  (see  Table  on  scientific  dissemination).  Three  selected  key  papers  are:  Cundy  AB,  Bardos  RP,  Church  A,  Puschenreiter  M,  Friesl-­‐Hanl  W,  Müller  I,  Neu  S,  Mench  M,   Witters   N,   Vangronsveld   J   (2013)   Developing   principles   of   sustainability   and  stakeholder  engagement   for  "gentle"  remediation  approaches:  The  European  context.   J  Environ  Manage  129,  283-­‐291    Kumpiene  J,  Bert  V,  Dimitriou  I,  Eriksson  J,  Friesl-­‐Hanl  W,  Galazka  R,  Herzig  R,  Janssen  J,  Kidd  P,  Mench  M,  Müller  I,  Neu  S,  Oustriere  N,  Puschenreiter  M,  Renella  G,  Roumier  P-­‐H,  Siebielec   G,   Vangronsveld   J,   Manier   N   (2014)   Selecting   chemical   and   ecotoxicological  test  batteries  for  risk  assessment  of  trace  element-­‐contaminated  soils  (phyto)managed  by  gentle  remediation  options  (GRO).  Sci  Tot  Environ  496,  510-­‐522.  Kidd  P,  Mench  M,  Álvarez-­‐López  V,  Bert  V,  Dimitriou  I,  Friesl-­‐Hanl  W,  Herzig  R,  Janssen  JO,  Kolbas  A,  Müller  I,  Neu  S,  Renella  G,  Ruttens  A,  Vangronsveld  J,  Puschenreiter  M  2014.  Agronomic   practices   for   improving   gentle   remediation   of   trace-­‐element-­‐contaminated  soils.  Int  J  Phytorem  (DOI:10.1080/15226514.2014.1003788,  in  press)    

 The  address  of  the  project  public  website,  if  applicable  as  well  as  relevant  contact  details    www.greenland-­‐project.eu    

1 Appendix to WP1 scientific report – Greenland Final Report

GREENLAND FINAL REPORT

Appendix to: Description of the main S&T results/foregrounds The GREENLAND network of field sites is a cross-European network of metal(loid)-contaminated sites where the efficiency of phytomanagement strategies has been investigated on a medium- to long-term, under various contaminant (trace element) types and loadings and soil and climatic conditions, with various plant species and cultivars. Task 1.1. Remediation option appraisal The general scheme for the remediation and phytomanagement of trace-element contaminated soils (TECS) comprises four stages: (1) risk assessment, (2) option appraisal, (3) implementation of remediation strategy and (4) phytomanagement (including biomonitoring and maintenance). Nine partners (INRA, CSIC, AIT, HAU, SLU, LfULG, Ineris, and IUNG) were deploying Gentle Remediation Options (GRO) at field scale for research and demonstration purposes. Five main types of historically contaminated sites were investigated, under different climatic and soil conditions, with either diffuse contamination on a large area (generally with agricultural soils) or local contaminations at mining sites, industrial facilities and landfills (with technosols at several sites) (Tab. A1). Case studies were categorized based on soil contamination levels (Kumpiene et al 2014) and exposome types which encompasses life-course environmental exposures for a biological receptor (Wild 20121). Based on the WP3 plant tests, the initial phytotoxicity of untreated soils followed the increasing order for dwarf bean: Bettwiesen, Freiberg, Piekary I & II < Lommel, Biogeco, Arnoldstein < Högbytorp, and for lettuce: Biogeco, Freiberg, Bettwiesen < Lommel, Piekary I & II, Arnoldstein < Högbytorp. Accordingly, the phytotoxicity of Bettwiesen, Freiberg and Piekary topsoils did not reflect their high total soil TE. Five exposome patterns were addressed based on main contaminants involved in pollutant linkages: Cu (Touro, Biogeco), Cd/Zn (Pb) (Lommel, Bettwiesen, Phytosed, Phytoagglo, Arnoldstein, Piedrafita, Högbytorp, Freiberg), As (Reppel, Jales, Freiberg), Cr/Mo (Rive de Gier), and metal/PAHs (Biogeco, Chaban-Delmas, Phytoagglo, Borifer). Tab. A1 Summary of the Greenland network investigating GRO in long-term field trials Partner Sites Country Sources GRO Main contaminants

Landfill PT-F Bettwiesen Switzerland former hot dip 1 Cd, Zn, Cr, Cu, Pb Zn galvanizing plant IUNG Piekary Poland Cd/Zn/Pb tailings 3 Zn, Cd, Pb INRA Chateauneuf France steel mill wastes 3 Cr, Mo, Zn, Ni, Cu, Pb

Atmospheric depositions on a large agricultural area HAU Lommel Belgium Zn/Pb smelter 1 Cd, Zn, Pb AIT Arnoldstein Austria Zn/Pb smelter 3 As, Cd, Zn, Pb LfULG Freiberg-Halsbrücke Germany Zn/Pb smelter 1, 2 As, Cd, Pb HAU/INRA Reppel Belgium/France As refinery 1 As, Zn, Pb, Cd

Wood preservation facility INRA Biogeco France wood preservative 1, 2, 4 Cu, Cu/PAHs

Mine, tailings CSIC Touro Spain Cu mining 1, 2 Cu CSIC Piedrafita, Rubiais Spain Pb/Zn mining 1 Zn, Pb, Cd INRA Jales Portugal/France Au mining 2, 3 As, Zn

Technosols, other sources SLU Högbytorp French trial Sweden Irrigation with landfill 1 (Cd), Cr, Zn leachates INERIS Phytosed ech 1 France dredged sediments 1, 2 Zn, Cd, Cu, Cr, Pb, Mo INERIS Phytagglo France brownfield 1, 4 Zn, Cd, PAHs INRA Chaban Delmas France embankments, harbor 2, 4 Zn, Pb, Cu, Cd, Ni, As, Hg, PAHs facilities                                                                                                                          1  Wild CP 2012. The exposome: from concept to utility. International Journal of Epidemiology 41: 24–32  

2 Appendix to WP1 scientific report – Greenland Final Report

INRA Borifer France metal surfacing 2, 4 Zn, Pb, Cu, Ni, As, Hg, PAHs ________________________________________________________________________________________________________________________________ 1: (aided) phytoextraction, 2: (aided) phytostabilization, 3: in situ immobilization/phytoexclusion, 4: rhizodegradation 1. Setting of conceptual models

Conceptual models (CM) were built for all sites (n=16), with information gained on main sources of soil contamination, soil exposome and ecotoxicity, biological receptors such as plant and microbial communities, initial pollutant linkages and risks on site and nearby. The procedure was derived from those reported by the Environment agency (UK). Exposures, organization/biodiversity and functioning of plant, animal, mesofauna and microbial communities were mostly taken into account, in line with exposures of animal and human populations (Fig. A1). At some sites (e.g. Biogeco, Jales, Reppel, Freiberg, and Piekary), vertical migration from the topsoil as well as wind erosion were considered. Clusters with different soil exposome and ecotoxicity were defined at many sites (e.g. Biogeco, Arnoldstein, Bettwiesen, Chaban-Delmas, Phytosed, Touro, and Piedrafita). Diverse physico-chemical parameters, soil and technosol types were of concern, from acid sandy soils to calcareous soils and alkaline technosol. Mixed contamination (trace elements and organic contaminants such as PAHs) was taken into account at 4 sites (Biogeco, Chaban-Delmas, Borifer, and Phytoagglo). Climatic conditions varied from cold climate and short summer (Sweden) to Mediterranean climate and dry summer (Spain).

Fig. A1 Conceptual model and pollutant linkages considered on site and nearby.

Different end land uses were considered: landscaping, recreation area, production of annual crops for (non-food) plant-based feedstock and biosourced chemistry, production of metal-excluder crops (grasses, cereals), production of wood from short rotation coppice (SRC) and tree planting in line with eco-technologies. Conceptual models were summarized in the datasheets for all case studies. 2. Risk assessment – pollutant linkages

Spatial distribution of soil contaminants, physico-chemical parameters, exposome (notably labile TE pools quantified by either single extractions or DGT) and phytotoxicity were characterized (in coordination with WP3) at most sites. The test battery for assessing soil exposome included generally standard methods (aqua regia, 0.1M NaNO3, 1M NH4NO3, and 0.05 M EDTA BCR), soil pore water (Rhizon MOM) and DGT. Initial and residual risks and pollutant linkages were quantified, in line with the GRO implementation and their (bio)monitoring. Impacts on soil microorganisms and plants were evaluated ex situ at most sites in coordination with WP3. Modelling of TE exposure vs. plant responses was done for metals (e.g. Cu, Zn, and Cd) and As based on total soil content and extractable fractions, soil pore water and the DGT method. The TE concentrations in plant products were determined at each harvest for all sites, and the potential uses of such phytomanagement-borne biomass were compared with the legislation, common values, and the biomass needs of local conversion chains. Concentrations of TE in leachates from the topsoil in either outdoor lysimeters or column, changes in labile TE pools in the soils and residual risks for plants and microbial communities were generally monitored over four years. To improve the detailed risk assessment several additional data were determined on site such as organization of plant communities (e.g. species richness, Shannon index) and interspecific variability of TE concentrations in plant parts. 3. Option appraisal

3 Appendix to WP1 scientific report – Greenland Final Report

The option appraisal stage aims at establishing which remediation option, or combination of options, can alleviate all pollutant linkages that present an unacceptable risk at the site (Environment Agency 2005, SUMATECS 2009). It includes: identification of feasible remediation options, detailed evaluation of options, and developing the remediation strategy. The Greenland project added another aim: the implemented GRO should improve the ecosystem services, notably provisioning services through biomass production for the bio-economy and other ecosystem services such as carbon sequestration, recycling of organic matters, water filtration, quenching of soil erosion and restoration of plant-microbe communities, without generating wastes and pollutant linkages. For the Greenland sites, the main concerns were to ensure that: • remediation option criteria selected for the soil are protective for controlled waters, plant, microbe and animal communities. (here remediation criteria were not based on total soil TE, but on either current legislation for labile (extractable) TE fraction in the soil, such as in Switzerland and Germany, for forages and feedstuffs (e.g. JORF 2003 in France), for foodstuff (EU Directive on Cd and Pb), for groundwater or on upper critical threshold values according to experts for the soil exposome in order to alleviate pollutant linkages). • relevant GRO were mainly selected for improving the biomass production and ecosystem services and reducing most pollutant linkages in line with the pathway soil - soil solution – plant, microbe, and animal communities; at some sites the TE bioaccessibility (through soil ingestion) was considered. • the Remediation Strategy addresses all (or most) relevant pollutant linkages • in general, requirements for waste management licences, environmental permits, discharge consents etc. were taken into account but not necessary for GRO selected in the option appraisal. To better determine the benefits and limits of feasible GRO for some or all clusters at one site, according to the selected conceptual scheme and end land use, it is recommended to compare them with the best relevant conventional remediation options, in parallel in pot and field experiments (with similar soil contamination). In case of failures of GRO in the long-term, the other remediation options would be deployed on the site clusters. Site-specific factors determining the appropriate GRO, i.e. nature of the conceptual scheme and risk management, location of treatable contaminants, overall strategy and implementation, and general criteria related to site and contaminants were addressed. Criteria related to technical basis, legal, and financial factors affecting the decision-making process such as engaging with stakeholders, were taken into account in the WP5. Depending on the sites, data for option appraisal were either previously published or produced by the project. Additional soil amendments and plant materials were investigated in WP4 in coordination with WP1 (e.g. biochars derived from either straw, pine bark chips, poplar twigs from Greenland phytomanaged sites or poultry manure, separately and in combination with zerovalent iron grit, compost; and red muds). Some of them (e.g. Linz-Donawitz slags) were further tested in field plots. For Cu-contaminated soils, biochar derived from poultry manure had a negative effect on plant growth whereas the C-Cure biochar combined with compost had positive effects. Several plant species were assessed to develop GRO at the Greenland sites (e.g. Miscanthus sinensis, Noccaea caerulescens, Arundo donax, Cana x generalis) (in coordination with WP4). Three main GRO were considered in the remediation strategies developed at field scale: in situ immobilisation/phytoexclusion, (aided) phytostabilisation, and (aided) phytoextraction (Tab A2). Option appraisals resulted in the selection of: ● phytoextraction at 13 field trials (5 with SRC, 5 with high annual biomass crops HBC, and 3 with hyperaccumulators), ● (aided) phytostabilisation: 11 field trials (3 with SRC and 8 with perennial herbaceous plants) ● in situ immobilization/phytoexclusion: 5 field trials. Tab A2: Option appraisals for several sites Option appraisals for several sites Remediation option Sites Feasibility Biochar derived from pine bark chips Chaban-Delmas, Borifer + Biochar derived from poultry manure Chaban-Delmas - Biochar derived from pine bark chips and compost Biogeco + Biochar derived from phytomanaged poplar wood Biogeco + C-Cure Biochar and compost Biogeco + Biochar derived from pine bark chips and iron grit Arnoldstein

4 Appendix to WP1 scientific report – Greenland Final Report

Biochar derived from poultry manure and iron grit Arnoldstein - Biochar derived from poplar wood and compost Biogeco + CaCO3 – reagent grade Piekary - Compost (GWDA municipal green waste and sludge) Piekary - Drinking water residue (DWR) +GWDA Piekary ++ Ca-phosphate – CaHPO4 +DWR + GWDA Piekary ++ Thomas basic slag (TBS) +GWDA Piekary + Linz-Donawitz slag (LDS) +Z + GWDA Piekary + Gravel sludge (GS) Piekary - Siderite (SID) – iron carbonate Piekary - Cyclonic ashes (CA) +Z + GWDA Piekary + Zerovalent Iron grit (Z) Piekary - Task 1.2. Implementation of remediation strategies The Greenland network is currently the frontrunner for the Northern hemisphere and an efficient EU tool, based on the 12thICOBTE and Conferences of the International Phytotechnologies Society (Syracuse 2013 and Heraklion 2014). Datasheets and success stories summarize the key information for each site. Implementation of the remediation strategy and demonstrating that it is and will continue to be effective. All GRO were implemented after the Tier 1 (Risk assessment) and most GRO were selected based on pot and/or mesocosm experiments (Tier 2, option appraisal). The Tier 2 was completed by the DST outcome from the WP5. In rare cases, GRO were implemented based on Tier 1, literature and information obtained at other sites (e.g. aided phytostabilisation at Phytosed, Touro). Licences were generally not needed as partners were working under the umbrella of local and/or national authorities, without the elimination or production of wastes and by-products. Main lessons gained on GRO implementation were: ● determine the areas of concern: field experiments were implemented on clusters defined by the initial risk assessment, identified pollutant linkages, and current/future land use. It was crucial to quantify the spatial variability of parameters driving the choice of feasible GRO according to the current/future land uses (for each cluster) and the related target/trigger values (notably those from the legislation and exposome) and other drivers (land value, time constraints, etc.). These parameters were (non-exhaustive list): total and labile pool for contaminants (when possible, including their chemical speciation) in the soil and soil pore water (if possible in the soil profile), capacity to buffer/resupply the soil solution, leachability, basic physico-chemical properties, texture/composition (define the soil type), and ecotoxicity of the (solid/liquid) matrices, climatic conditions including water supply and its annual distribution, etc. A key point is to survey the water supply and requirements by different plants. ●  account for any specific requirements related to the selected feasible GRO and the best conventional option (to be compared). Spatial variability of pollutant linkages is a pivotal parameter. Before implementing field experiments for testing selected GRO, attention must be paid to the plant communities already colonizing the site/clusters (if any). Presence and habitats of animals (including insects, soil mesofauna, etc.), the slope and the terrain relief in general must be recorded. Information was gathered for most Greenland sites on the spatial variability of pollutants and their linkages, plant candidates for GRO, and eventually (native) plant populations and associated microbes, which can be used directly or selected to obtain efficient partnerships. Sub-site(s) were defined at several sites (e.g. Touro, Biogeco, Phytosed, Arnoldstein, etc.) allowing to statistically exploit the field plots. The spatial variability of soil ecotoxicity was assessed for each cluster (at least a plant test with a sensitive plant species and an indicator of pollutant linkage or exposure such as the NH4NO3-extractable soil fraction). ●  compare the best conventional technology(ies) in parallel with the selected GRO emerging from option appraisal. In case of GRO failure, the conventional technology will be an immediate alternative. To better assess the benefits/limits of the GRO’s, it is better to compare with the best conventional technology to provide relevant and convincing information to the landowner and the stakeholder core. ● don’t upscale directly from ‘pot experiments’ to ‘full-scale’ (in situ) deployment on the cluster(s) without the return skill of biomonitoring and maintenance for several years. Field plots must be tested on a long-term, especially to address and optimize some aspects that are difficult to investigate with potted soils: e.g. variability of climatic conditions, colonization by animal communities, pests, ageing of soil amendments, extension of the root systems, etc. For tree management, enough space must be allowed between the plots, as root systems can extend horizontally more than 10 m for poplars as well as the shading effect. As far as vertical migration to the subsoil and groundwater is of concern, try to establish an in situ lysimeter system (even a basic one with containers) or an in situ leachate sampling system underneath the plots to long-term assess the quality and the ecotoxicity of the leachates. Horizontal migration of the contaminants through wind erosion and other natural agents (water runoff), in particular to inland water and allotments, must be considered too. ● fencing: A single fence around the whole site may be necessary (notably to restrict the entrance) but it is generally not sufficient to prevent potential damages caused by the mammal herbivores (i.e. rabbits, field rats, deers, etc.). It should be

5 Appendix to WP1 scientific report – Greenland Final Report

complemented by fences around small clusters (especially at the start of the phytomanagement, to protect the trees and other attractive plant species; individual fences around trees are less time-consuming but their efficiency is lower. ● plot size: define reasonable plot size for avoiding the edge effects and permitting a long-term (>5 years) monitoring and maintenance, notably for soil and plant samplings. Set up field plots according to the spatial variability of parameters listed above; pay attention to allow sufficient space between the various options; always remember that tree roots, and its associated hyphosphere, can sense better conditions over more than 10-15 m; pay attention to the shading effect which may occur with the canopy development. In case of slopes, the common technique of terraces can be used as well as fiber nets to counteract the soil run-off till the establishment of the vegetation cover, as for to vegetate ski tracks. Starting from seeds, some light mulch (with straw, fern fronds, bark chips, coconut nets, etc.) to trap the seeds can be necessary (and avoid migration with natural agents or bird predation). ● don’t forget to monitor the foliar exposure: at some sites, local emissions, atmospheric fallout and windblown dust may occur, contributing to foliar exposure. Place some pots with uncontaminated soil to grow some plants, e.g. grassy crops and young trees, for quantifying such foliar exposure. For comparison, potted contaminated soils under remediation must be placed at an uncontaminated site without relevant emissions or under controlled ‘clean’ conditions. ● adopt appropriate agronomic practices: GROs are essentially based on ecology, microbiology, ecotoxicology, and biogeochemistry, and their success will inevitably depend upon the careful implementation of effective agronomic practices such as crop selection, crop rotations, intercropping, planting density, fertilization, irrigation schemes, bioaugmentation, weed, pest and herbivory management (Kidd et al 2015). Conventional agricultural methods can be modified so as to suit both the characteristics of contaminated soils, and to meet the requirements of effective phytoremediating crops. Agricultural practices can be incorporated into GRO as a means of optimizing metal(loid) extraction, immobilization or the prevention of their excessive transfer into the food chain (without inducing TE and (macro)nutrient deficiencies). In addition, these practices can improve plant biomass production, nutritive status or pest management. Sometime during winter, the flooding of soils may prevent the harvest machines to enter in the field. This can be a bottleneck for harvest of SRC, Miscanthus, etc. ● Implementation of plant species: for phytomanagement, the choice of initial plant/microbe partnerships must account for the local conversion chains of biomass (generally the biomass production on one rather small site is not enough to financially support a dedicated local valorization plan; this biomass must be commonly merged with similar biomass from other sites (forest, SRC, agricultural field, green wastes, etc), provided that their composition is suitable with the process or its marketing image. Phytotoxicity and other stress factors can limit the performance of the plant species used in the GRO. The careful selection of plant species and optimization of growth are key elements in successful phytomanagement of TECS under different pedo-climatic conditions. Plants must not only show tolerance to the contaminant(s) present but also resist other abiotic and biotic potential stresses, e.g. water stress, soil acidity, frost, soil erosion/compaction, herbivory, pests, nutrient deficiency, salinity, etc. A source of TE-tolerant plant genotypes is the pioneer vegetation colonizing contaminated sites or present nearby. Screening and selection of TE-tolerant plant species and genotypes (e.g. Agrostis sp., poplars and willows, tobacco, etc.) were made in WP4 for application under real field conditions. Piekary site provided information on long-term performance of various grass species on tailings reclaimed with biosolids. Within the same plant species different ecotypes, cultivars, varieties and clones can differ in their response to contaminant excess. While tolerance to contaminants in question will always be vital, at other times the selected plant will depend on the GRO to be used e.g. TE-accumulating plants (phytoextraction), TE-excluding plants or crop species (phytostabilisation/phytoexclusion). Some herbaceous and woody Fabaceae can be included in crop rotation or in mix stand to promote the nitrogen supply by fixation of atmospheric nitrogen. - for woody crops: competition between young trees and the herbaceous plant communities (notably grassy crops) can be adverse for tree development. Try to implement the young trees before to implement the herbaceous crops underneath and in between, if there is a need to increase the vegetation cover and reduce the contaminant migration through natural agents. It is pivotal to irrigate trees in year 1 (and sometime year 2) during dry periods to increase the survival rate and promote the establishment of their root systems (depending on soil type, climatic conditions, etc.). Pay attention to the slope, potential soil erosion and/or flooding. In case of excluder-based SRC for bioenergy purposes, the selection of genotypes can be based on their characteristics in line to conversion processes, e.g. calorific value, bulk density, moisture content, ash and extractive content. Transplantation of mycorrhizal trees was more successful than that of non-mycorrhizal ones and the on site mycorrhization of tree cuttings. Inoculation of native TE-tolerant ectomycorhizes can initiate a synergetic fungi succession. Salix and Populus clones show high variations in biomass production, TE tolerance and accumulation patterns in roots, leaves, and even in wood between clones. Some species and clones of willow have high bioconcentration factors (BCFs) for Cd (up to 27) and Zn (up to 3). Given the ample variation in metal accumulation, best-performing clones can be selected based on their TE-tolerance, uptake efficiency (accumulating clones for phytoextraction vs. excluding clones for phytostabilisation), translocation from roots to shoots, and biomass production. Clones can be selected for their ability to accumulate certain metals (e.g. Cd and Zn) while at the same time immobilizing elements such as Cu or Pb. Evidences of tolerance to TE and fungal and insect infection, e.g. leaf rust (Melampsora sp.) and lace bug (Monosteira unicostata Muls. and Rey), cold and drought adaptation were revealed at the Lommel site.

6 Appendix to WP1 scientific report – Greenland Final Report

● combine phytomanagement and ecology: establish natural and passive habitats to host and promote reproduction of the biological auxiliaries (notably beneficial insects and birds) and counteract bioagressors. Think about the connection of clusters with the other ones nearby. Use corridors allowing the predators (fox, raptors, etc.) to hunt; these corridors can be combined with the access required for monitoring and sampling as well as the harvest machines. Avoid a full site monocultures to alleviate the selection of pest populations (e.g. use diverse clones/genotypes for trees in clusters; use a crop rotation in case of annual plants.  

● Phytomanagement can combine some GRO: The phenotype of plant species in response to TE excess is element dependent and a plant assemblage can support various GRO at the same time on mixed-contaminated soils. For example a poplar SRC can simultaneously phytostabilize Cu/Pb in its root system, phytoextract Cd/Zn in its aerial parts and promote the rhizodegradation of xenobiotic organic compounds.

Task 1.3. GRO implemented and biomass production 1.3.1. In situ immobilization/phytoexclusion This GRO can be implemented as either a long-lasting (phyto)management option or a temporary, reversible one that can be later modified based on the monitoring results from the phytomanaged plots. Decreasing the labile TE pools in TECS by incorporation of soil conditioners and the use of excluder plants are both main approaches. Different soil conditioners were investigated on a long-term and at field scale, i.e. phosphates, composts and technosols, iron bearing materials (iron grit, gravel sludge), and alkaline materials such as alumino-silicate slags, marl lime, biosolids, and dolomitic limestone. - Organic matter: composts are frequent in the amendment combination for promoting crop production and soil quality. Their quality (C/N ratio, seed bank, labile P pool, etc.) is pivotal. They were used, singly and in combination (Biogeco) and compared to technosol mixtures (Touro), especially in Cu-contaminated soils. Compost was more efficient to promote the vegetation cover at Touro. Biosolids can be successful and economically viable option when large volumes of amendment are needed (smelter tailings). Caution must be paid when using rather “fresh” organic material in case of a labile pool of Cu, Pb, As, Mo, Cr, Sb, and Sn. Dissolved organic matter (DOM) may transiently increase the soluble complexed (for metals) or free anion (for metalloids) fraction in the root-zone with consequences on TE leaching when plants were not able to fully use the leaching water and take up soluble TE. - alkaline materials, through changes in soil pH influenced physico-chemical and biological reactions into the TECS, with consequences on the chemical speciation, sorption by bearing phases and mobility of TE (Biogeco, Phytosed, Freiberg, Piekary). Over-liming may however induce nutrient deficiency and mobilize TE in oxyanion forms. Marl lime (Freiberg), dolomitic limestone (Biogeco), biosolid and by-product lime (Piekary), and alumino-silicate slags such as Linz-Donawitz slags (LDS), with and without P spikes (Biogeco, Phytosed) were assessed. - other soil conditioners: Fe/Mn bearing materials such as zerovalent iron grit (Z) and water treatment sludge (WTS), gravel sludge, red mud and siderite bearing material were tested singly and in combination notably with compost. All were tested in field plots. Incorporation of Z and LDS into TECS (Biogeco) was split to avoid the pepite formation and better homogenize the amended soil. Various incorporation ways (injection, tillage, slurry) were assessed at Arnoldstein. End land use: annual crop production Staple crops and oilseeds: Cultivars within species from major staple crops such as wheat, barley, rice, potato and maize differ widely in their ability to accumulate metal(loid)s. Selection of efficient excluder cultivars for cultivation on contaminated and remediated land contributes towards reducing the entrance of non-essential TE, and also avoiding the excess of essential ones, into the food chain. Cd is of highest concern regarding metal uptake into the food chain as well as As, Mo, Se, Tl and Hg. Selection of the most appropriate cultivars for use on TECS can ensure that food and forage production is in compliance with the respective regulations on threshold TE contents. For example, The Operating Company for the Environment and Agriculture of the Saxon State (Germany), the Austrian Agency for Health and Safety (Austria) and Arvalis (France) are testing TE uptake behaviour of currently available cultivars in both field and parallel batch experiments. However, in many countries, farmers often have limited access to excluder type cultivars on a regional base due to the lack of information about the uptake properties of available cultivars. Moreover, since commercial availability of certain cultivars changes rapidly, the data for current cultivars has to be frequently updated to allow adequate selection of cultivars appropriate for contaminated land. Excluder maize, barley and potatoes cultivars were long-term assessed at Arnoldstein. Use of the excluder-phenotype Bodega vs. accumulator-phenotype Hellana reduced barley grain Cd by over 40 %. In combination with the incorporation of gravel sludge and red mud into the contaminated soil, a further >30 % Cd uptake could be avoided. After five years,

7 Appendix to WP1 scientific report – Greenland Final Report

soil amendments at Arnoldstein were still effective immobilizing agents illustrating the efficiency and cost-effectiveness of in situ stabilization and phytoexclusion. Management practices recommended by the authorities to the owners of contaminated land in the Freiberg area were to increase soil pH to values of 5.8 to 6.5, moderate phosphorus (P)-fertilization at the beginning of the growing season, increase redox potential, and use harvesting methods which minimize contamination of grass forage by soil particles. Consequently crop rotation including winter oilseed rape, winter wheat and spring barley at Freiberg-Hilbersdorf in combination with marl lime application. Soil pH at Freiberg-Hilbersdorf was slightly changed by lime and P application and was generally increased in year 4, which resulted in a decrease of mobile Cd (by 50-75%). Based on BCF of Cd, Zn and Pb, the barley cultivar “Salome” was shown to less accumulate these metals in its grains than the Marthe cultivar. Based on EU directive 2002/32/EC (2002) these barley grains were suitable as single fodder. Considering changes in element transfer into plant parts as affected by amendment options, grain Pb differed between the control and P treatment with highest concentration and the limed treatments with low concentration, especially for the combined fertilized treatment. The biomass production of winter oilseed rape for both cultivars was within the common range of yields for this German region (2.4 – 4.4 t/ha). Those of winter wheat were below the range (5 – 8 t/ha), especially for the low accumulating cultivar Türkis, which produced a lower grain yield than to the high accumulating cultivar Tiger. The grain yields of spring barley were below the common range (4.2 – 7.4 t/ha) with slight differences between both cultivars. Grassland management: Grassland based on TE excluder grassy crops is one relevant GRO to alleviate windblown dust and water runoff on large TE-contaminated areas, notably with low fertility. At Arnoldstein, shoot DW yield reached 5 t/ha/yr. The most efficient soil conditioner (gravel sludge and red mud, slurry management) was reducing the labile pools of Zn (-90%), Cd (-80%), and Pb (-90%) in the soil. Plant monitoring based on Plantago lanceolata indicated reduced shoot concentrations for Cd (-70%) and Zn (-77%). Shoot Cd and Pb concentrations of harvested grass mixture just exceeded the maximum permitted concentrations (MPC) in forages. At Freiberg-Hilbersdorf, last marl lime application was made in autumn 2013 and last soil and plant samples were taken in May 2014 from the grassland trial. Soil pH (CaCl2) varied between 4.3 and 5 in the unamended soil. It reached pH 6 at 2t marl lime/ha and 6.5 at 4t/ha. Consequently extractable Cd in the soil was reduced from 0.6-0.7 to 0.05-0.1 mg/kg soil DW. Grass shoot DW yield varied from 2 to 5.75 t DW/ha depending on season (3 cuts/yr) and soil amendments, which high values in May and July and low ones in September. It was enhanced in May by marl lime application at 4t/ha. Shoot As concentrations of grass did not differ between unamended and 2t/ha-amended soils on the 2012-2013 period and was in the 0.25-0.5 mg/kg DW range. It started to decrease on the third year. Shoot As concentration was higher in the 4t/ha-treated soil, reaching 1.5-1.75 mg As/kg DW in 2012-2013, despite high shoot DW yield, and also decreased to 0.6 mg As/kg DW in 2014 after the last marl lime application. Shoot Cd concentration ranged between 1-1.5 mg/kg DW in 2012 and did not differ across the treatments. It decreased in all treatments in 2014 (0.5-0.8 mg/kg DW), but lower values in marl lime-treated soils were statistically similar to the unamended soil. Shoot Pb concentration varied from 0.3 to 4 mg Pb/kg DW in average and reached 8 mg/kg DW in some shoot samples from the 4t/ha-amended plots. It was decreased in year 3 for all plots with a lower value at 2 t/ha (0.3 mg Pb/kg DW) compared to the unamended soil.

At Piekary (PL): The grass mixture consisted of local cultivars: Festuca rubra L. cv. Atra, Poa pratensis L. cv. Alicja, Festuca arundinacea Schreb. cv. SZD, and Festuca ovina L. cv. Sima. 17 years after biosolid incorporation, water-soluble fractions of major contaminants (Zn, Cd, and Pb) in the soils remained at low levels, in line with soil pH and Ca-carbonate distribution over the field. Soil bacterial communities were highly diversified in amended soils. Dehydrogenases activity increased as water extractable metal (Cd, Zn) fractions in the soils were reduced. Plant cover and biomass production depended on the soil treatment being highest soils amended with biosolid combined with by-product lime. Untreated tailings outside the reclamation area remained barren. The plant cover is not managed and plant community organizes itself. At the field where grass species were tested, the most persistent grass species were Poa pratensis, Agrostis capillaris and Festuca ovina. These species covered the largest area of the field 17 years after remediation among all grasses. A substantial part of the areas was covered by colonists - Calmagrostis epigejos, Hypochoeris radicata, Melandrium album, Artemisia vulgaris, Daucus carota and Solidago gigantean. 1.3.2. (aided) phytostabilisation

8 Appendix to WP1 scientific report – Greenland Final Report

Two related GRO were assessed: ● phytostabilisation with various plant covers, i.e. mycorrhizal and non-mycorrhizal trees, and perennial grasses, matrices (soils, tailings), climatic conditions, strategies and socio-economic opportunities. ● aided phytostabilisation combining TE in-situ stabilization, mainly through changes in sorption processes and pH in the soils, and phytostabilisation, i.e. (non-)mycorrhizal trees, perennial grasses. The purposes were (1) to cultivate trees and/or perennial grasses on TECS, (2) to manage trees as Short Rotation Coppice (SRC) or fast growth plantation, and (3) to decrease labile TE pool in the root zone. 1.3.2.1. SRC with and without grass cover/herbaceous layer

SRC parameters: Many tree species are suited for phytostabilization due to their deep root systems, high transpiration rate, high TE tolerance, and ability to grow on nutrient-poor soils. Trees can stabilize less mobile metals (e.g. Cu, Pb, and As) in the soil by physically preventing migration, leaching, and soil dispersion; alternatively, they can immobilize TE through uptake and accumulation by the roots into the plant, adsorption on the root, and precipitation in the rhizosphere. At Biogeco (Cu-contaminated soils) two cuts were made for fertilized mycorrhizal poplars (in years 4 and 7), whereas minimum values of potential SRC biomass to initiate the harvest of willows and non-mycorrhizal poplars were not reached. For willow SRC, it can be done in year 9 only for ectomycorrhizal trees. Shoot DW yield of poplar SRC varied from 20 to 270 t DW/ha showing the spatial variability of soil exposome, fertility and water supply, 135 t DW/ha was even reached in some untreated plots nearby other fertilized plots managed by phytoextraction. In year 6 after amendment incorporation into the Cu-contaminated soil, compost (OM) increased poplar growth compared to the untreated soil (UNT), whereas addition of dolomitic limestone (DL) resulted in less significant increases. Both OM and OMDL promoted the growth of Amopha fruticosa. Effect of soil amendment was not significant in the long-term for both willow species. At Phytosed, after six months, grass lines were mechanically removed and replaced by a tarpaulin at the expected willow place. Two willow cultivars (Tordis and Inger) were planted (12,000 willows ha-1) in SRC for the biomass production. B. cespitosa and the natural colonizers were mowed to maintain as low as possible the competition for water and nutrients with the willows. The survival rate of willows in year 2 was 89% accounting for all plots, but it dropped to 75% for Tordis in several amended plots and the rate of chlorotic leaves reached 30-50% in these plots. Consequences of over-liming and Cr/Mo excess were hypothesized. The grass may also compete with willows. The foliar Cd concentrations of willows were high (10-30 mg Cd/kg), the Tordis willow clone showing higher values than the Inger one. This difference in foliar Cd concentration between the willow clones was observed both in the Thomas basic slag (TBS)-amended plots and in the control plots. Values in year 1 were far higher than frequent concentrations in willow leaves from uncontaminated soils (<2 mg kg-1 DW). Similarly foliar Zn concentrations ranged from ~ 2000 to 3500 mg kg-1 DW whereas common values varied from 81 to 296 mg kg-1 DW. The alkaline amendment did not decrease foliar TE concentrations of willows. Wood and bark Zn and Cd concentrations in year 2 (2 mg Cd/kg DW) were lower than those in willow leaves. Bark concentrations (10-15 mg Cd/kg) were higher than wood concentrations and concentrations increased with the height of willow due to the increase of bark proportion. Tordis willows accumulated more Cd than Inger, in accordance with the leaf results. Compared to initial Cd and Zn concentrations in both willow cultivars before plantation (~ 2 and 150 mg kg-1 DW in wood with bark, respectively), the Cd/Zn concentrations increased after 2 years. The alkaline amendment did not reduce or at least stabilize the TE concentrations in aerial plant parts of willows. Herbaceous layer: The sediment landfill site at Phytosed (FR) is contaminated by TE, mainly Zn and Cd. A commercial alkaline by-product of steel industry used in agriculture (Optiscor) was incorporated in September 2011 (rate 9 t ha-1 to optimize both metal immobilization and willow growth, pH 8) into the technosol for reducing the metal mobility and promoting the grassy crop. Barchampsia cespitosa was used as a plant cover to reduce vertical and horizontal TE transfers, and was expected to alleviate the propagation of Fallopia japonica, an invasive species colonizing the technosol. After 2 years, the vegetation cover roughly reached 100% and foliar Cd concentrations in B. cespitosa were lower than 0.5 mg kg-1 DW and approximated common values (0.05-0.2 mg kg-1) for grasses grown on uncontaminated soils. Despite the high total soil Cd, this grassy biomass was suitable for composting. Cd foliar concentration was even reduced in year 2. Averaged foliar Zn concentrations ranged from 180 to 270 mg kg-1 which

9 Appendix to WP1 scientific report – Greenland Final Report

overlapped both common (27-150 mg kg-1) and upper critical threshold (100-400 mg kg-1) values in grass shoots from uncontaminated soils. No phytotoxicity symptoms were observed on aerial plant parts. Concerning the mineral amendment, its efficiency was not demonstrated neither on foliar TE concentrations nor extractable TE concentrations in the technosol. Indeed, no reduction in concentrations was observed. The commercial cultivar, B. cespitosa, is a good candidate for phytostabilisation (i.e. success of the plant cover, tolerance to the technosol conditions, shoot TE concentrations close to common values for grasses on uncontaminated soil). This grass competes well against the invasive species (beneficial effect of phytostabilisation; see previous report). Until now, the selected soil amendment did not succeed. Future work will address the expected mechanisms (speciation, OM, CaCO3 stock, etc.). In this case, the combination of aided phytostabilisation using a grass cover with the plantation of willows to produce biomass for bioenergy is not successful (i.e. grass and willow competition for water and nutrients, sensitivity of the selected willow clones to the labile pool of contaminants and other factors such as willow leaf beetle, herbivores..., generation of costs rather than economic benefits [see WP5 for economic data]). One alternative would be to put the grass several years after the willow plantation to avoid the grass competition. This option poses the following questions: is it technically feasible? What about the risks in this case? Replace grass by mulch? Is it economically viable? Is it possible to find other fast growing trees (than willows and poplars) or cultivars with very low TE accumulation? In this case study, benefits of biomass production do not compensate costs linked to set up and monitoring of both aided phytostabilisation and willow plantation. This result questions the possibility to decrease these costs. This could be achieved by recalculating cost and benefits with other protocols. Responses to bioagressors: Both willow cultivars Inger and Tora at Phytosed were susceptible to the imported willow leaf beetle (Plagiodera versicolora). A severe attack occurred in early spring 2014. Consequently an organic insecticide (pyrethrin) was applied after the leaves have flattened out; this efficient treatment was leading to the leaf re-growth. A similar biotic interaction was occurring at Lommel. The use of native poplar and willow at Biogeco reduced disease incidence, particularly from Melampsora rust. Assemblages of plants and microbes Effect of ectomycorrhizae and fertilization: inoculation with mycorrhizal fungi and plant-associated bacteria (rhizobacteria and endophytes) may improve plant growth and modify soil metal mobility and their uptake/translocation by woody crops, notably in TE-contaminated soils and mine sites. At Biogeco, the maximum stem height of mycorrhizal trees was higher than that of non-mycorrhizal ones. However, poplars in plots nearby fertilized plots phytomanaged by phytoextraction have extended their root system and took advantage of these plots for their growth. Their root system was able to detect lower labile Cu pool and NPK supply in plots nearby. This underlines the influence of plot size and the interspaces for long-term assessment of SRC. Ectomycorrhizal poplars inoculated with Cu-tolerant endophytic bacteria were also obtained in greenhouse and then transplanted at Biogeco. Touro (ES): The mine tailings of the non-active Cu mine cover an area of approximately 550 ha. The implemented GRO involved establishing a short rotation coppicing system or a grass cover with the principal objective of reducing Cu mobility. The geological substrate is amphibolite, with significant quantities of metal sulphides (pyrite, pyrrhotite, and chalcopyrite). The mine-soils (Spolic Technosols (Episkeletic)) are characterized by their extreme acidity (pH 2.8-3.5), low C, N and P, and high concentrations of Cu (319-774 mg/kg). Cu contamination shows considerable heterogeneity across the site. Tailings were amended with three mixtures: composted municipal solid wastes (compost) and two technosol mixtures. Technosols were based on organic (anaerobic and aerobic sewage sludge) and inorganic wastes (aluminium oxides, iron oxides, fly ash from wood bark combustion, and foundry sand). Plots were planted with different metal-tolerant clones of Salix (S. caprea and S. viminalis) and Populus nigra, or with a grass cover Agrostis capillaris cv. Highland. Mortality was high on technosol-amended plots but low on compost-amended plots. Growth and survival (70-80%) of woody trees was optimal in compost-amended plots. After three years tree height was highest in S. viminalis and P. nigra (reaching up to 3-4 m). Changes in soil exposome / TE mobility in soils Touro: In year 3, soil NaNO3-extractable Cu concentrations remained low (<1 mg/kg) in all treated soils without influence of the vegetation cover type. Soil pH was 3.5 before GRO implementation, and in year 3 remained between 6.0 and 7.0 in compost-amended soils. Soil pH was higher in soils under Salix, followed by Agrostis and finally, unplanted soils.

10 Appendix to WP1 scientific report – Greenland Final Report

Phytosed: In year 2 extractable Zn and Cd fractions (roughly 0.4-0.8 mg Zn and 0.001-0.0015 mg Cd/kg soil) did not differ between the amended and unamended technosols. Biogeco: In year 5, Cu concentration in the soil pore water was higher in the compost-amended soils for both mycorrhizal and non-mycorrhizal trees and the lower in the limed soils with mycorrhizal trees. Restoration of soil microbial activity and communities: At Touro, soil enzyme activities (involved in C, N and P cycles) were monitored over time. In parallel, shifts in the structure of the soil bacterial community (total Eubacterial community, α- and β-proteobacteria, Actinobacteria and Streptomycetaceae) were compared over time using the Denaturing Gradient Gel Electrophoresis (DGGE) technique. Soil enzyme activities were lowest in untreated soils, and increased with time in amended soils. A plant-induced effect was also observed: activities were higher in plots planted with woody trees, followed by Agrostis, and lowest in unplanted plots. Similarities in DGGE fingerprints based on 16S rDNA amplified fragments were analysed. At each sampling period (after 1, 2 and 3 years) the similarity dendrograms showed a separation (similarity of <20%) between the bacterial community of soils sampled before GRO implementation (i.e. time=0) and that of phytomanaged soils. At all sampling periods (1-3 years) the soils amended with compost, technosol 1 or technosol 2 formed three distinct clusters of DGGE profiles. In year 1, within each of these three clusters there were clear separations corresponding with soils sampled from unplanted plots, or cultivated with Agrostis or Salix (similarities of <60%). With time the three amendments continue to cluster separately but there is some intermixing of DGGE profiles from either Agrostis or Salix cultivated plots. In general the soil bacterial communities continue to cluster separately according to the plant species. Similar patterns were also observed in the DGGE profiles of α- and β-Proteobacteria, Actinobacteria and Streptomycetaceae populations, albeit sometimes to a lesser extent than in the DGGE profiles of Eubacteria. Bacterial diversity increased in all phytomanaged soils compared to the untreated soils. There was a trend towards an increase in bacterial diversity with time, and also a higher bacterial diversity in planted soils (albeit Agrostis or Salix) compared to unplanted (but amended) soils. 1.3.2.2. Grassy crops (only)

TE excluder perennial herbaceous crops such as switchgrass (Panicum virgatum), miscanthus (Miscanthus spp.), giant reed (Arundo donax), and vetiver (Vetiveria zizanioides) have wide climatic adaptability, low production costs, suitability to marginal lands, relatively low water requirements, low nutrient and agrochemical needs, and potential environmental benefits (e.g. carbon storage through their deep and well-developed root system). They can provide feedstock for the energy sector or essential oils in Europe and North America. The invasiveness of some of these species (e.g. giant reed) is a controversial topic. They were assessed with potted Cu-contaminated soils. Thereafter, Miscanthus x giganteus and Vetiver were implemented at Biogeco. After four years, Miscanthus as Vetiver did not colonize other plots and nearby area. Co-cropping of leguminous species (Ornithopus compressus, Medicago arabica and Trifolium pratense)  with Miscanthus did not promote its shoot DW yield. For Vetiver in year 4, shoot DW yield potentially reached 38t/ha in uncontaminated plots, 7-15t/ha in the amended Cu-contaminated plots, and only 2.6 t/ha in the highly Cu-untreated plots, as total Cu in soil pore water increased from 0.2 to 0.9 mg Cu/L. It always demonstrated a Cu-excluder phenotype, shoot Cu concentration being in the 10-13 mg/kg range with no influence of soil Cu contamination, which alleviate potential herbivory exposure. For Miscanthus, shoot DW yield in year 3 varied from 0.07 (Unt) to 1.8 (OMDL) t DW/ha. Its shoot Cu concentration ranged from 7 (OMDL) to 95 mg Cu/kg DW (Unt) (values in year 1 were higher due to a lower shoot biomass: 16.6-507 mg Cu/kg), slightly over the Vetiver values. The single incorporation of compost and dolomitic limestone still reduced labile soil Cu and Miscanthus exposure in year 7. Change in labile soil Cu and in shoot DW yield (dilution effect) explained differences in shoot Cu concentration. Shoot Cu removals in year 3 varied from 3 to 17 g Cu/ha as shoot DW yield increased, depending on shoot Cu concentration and shoot DW yield, with maximum at median soil Cu contamination and minimum for soil Cu/PAH contamination (in line with foliar symptoms of N and water deficiencies). Biomass sorghum (Sorghum spp.) was assessed in potted soil and field plots (Biogeco). Both cultivars for biomass and bioenergy were not successful at field scale, being too sensitive to Cu excess and low water supply in sandy soils. Grassland: At Biogeco compost (OM) and dolomitic limestone (DL), singly and in combination (OMDL) were assessed in comparison with the untreated soil (Unt). In year 7, most plots initially planted with grasses were dominated by an assemblage of Cu-tolerant Agrostis capillaris and A. gigantea whatever the soil treatments. Other

11 Appendix to WP1 scientific report – Greenland Final Report

introduced grassy species such as Sporobolus tenacissimus and B. cespitosa are disappearing. Cytisus scoparius are colonizing the plots and their shoots are annually harvested to avoid the development of a bush canopy. Grass cover is declining in one block, may be due to competition with roots of poplars located nearby. Shoot DW yield was influenced by soil treatments: in year 7, values were higher in the compost-amended plots compared to the limed ones. Shoot Cu concentration was slightly lower for the grass species harvested in the OMDL plots. Highest shoot DW yields in the OM and OMDL plots led to maximum shoot Cu removals. Shoot Mg, K, Na, B and P concentrations were higher whereas shoot Zn and Al concentrations were lower for the OMDL plants compared to the Unt plants. In year 5, total Cu concentration in the soil pore water was increased in the compost-amended soils as compared to the limed soils. Chaban-Delmas (FR): A total of 72 plant species were identified in the grassland. 32 species were occasionally present, i.e. they always represented less than 1% of the vegetation cover on all subplots. For the other 38 species, three plant subsets were determined. The first one (Subset I) included 40 subplots located in the area center and was dominated by Medicago sativa and Lolium perenne, followed by Vulpia myuros, Holcus lanatus and Eleusine tristachya. This subset had the lowest plant species richness. The second subset (Subset II) including 15 subplots, located at the south part of the platform, was dominated by Arrhenatherum elatius, Bromus sterilis, Holcus lanatus and Dactylis glomerata. It displayed a median value for the species richness. Melilotus albus, Trifolium arvense, and Trifolium pratensis were dominant in the third subset (Subset III), located at the north of the platform and with the highest bare soil percentage. This one had the highest species richness. Touro (ES): Agrostis capillaris cv. Highland was successfully established in both compost- and technosol-amended plots. Shoot Cu concentrations were within normal levels for grass species growing in uncontaminated soils, and significantly lower than in grass species colonizing the surrounding untreated tailings. Nutrient concentrations were increased in all amended soils but particularly in technosol-amended plots. Grass biomass was harvested each year (2012-2014) and shoot DW yield was highest in compost-amended plots. 1.3.3. (Aided) phytoextraction The aims were to (1) quantify the biomass production, (2) the plant ionome (notably TE concentrations) and (3) the TE phytoextraction. Additional aims were (4) to improve agricultural practices, (5) to enhance ecosystem services such as C sequestration and microbial activities, and (6) to create economic opportunities from the biomass. Plants must be able to accumulate high TE concentrations in their harvested parts and have a reasonably high biomass production. Relevant options were TE-hyperaccumulators and secondary TE accumulators. Three main options were implemented to address three mains situations, Cd/Zn (Pb), Cu (Cu/PAH) and As/metal excess: high-yielding crops (HYC), short rotation coppice (SRC), and herbaceous hyperaccumulators, in monoculture and co-cropping. Influences of soil conditioners such as compost, Linz-Donawitz slags, soil acidifying agents (citric acid, S), co-cropping were investigated to enhance TE phytoextraction. 1.3.3.1. High-yielding crops (HYC)

High-yielding crops (annuals or perennials) are recognized as viable alternatives for TE phytoextraction (particularly Cd, Se and Zn) if they show relevant shoot TE removals (i.e. moderate-high BCF and high shoot yield). In vitro breeding (cell and callus tissue culturing on metal spiked media) and chemical mutagenesis can improve the metal tolerance and phytoextraction capacity of high-yielding crops such as tobacco and sunflower. These non-genetically modified plants can be tested under real field conditions without legal restrictions. Commercial sunflower cultivars accumulate only moderate metal concentrations, but their high biomass production makes them attractive for Cd/Zn phytoextraction. Some oleic cultivars can provide both relevant oilseed yield and shoot Cu removal. Chemical mutagenesis (EMS) was used to improve shoot metal concentrations and biomass production of a sunflower inbred line IBL04. At the Rafz site (Switzerland), shoot metal removals by the sunflower mutant were up to 7.5-, 9.2- and 8.2-fold higher for Cd, Zn and Pb than the inbred line, respectively. As monocultures can lead to a decline in biomass yield due to the depletion of nutrients, occurrence of diseases, pests, and weeds, and have a negative effect on soil fertility, crop rotations such as sunflower/tobacco (with winter fodder pea at Bettwiesen, and white clover at Biogeco as cover crops during winter for green manure and limiting soil erosion) were investigated. Fibre hemp (Cannabis sativa) and kenaf (Hibiscus cannabinus) were cultivated at Lommel.

12 Appendix to WP1 scientific report – Greenland Final Report

Tobacco and sunflower mother-clones and variants (from PT-F) were cultivated at 5 sites managed by INRA, CSIC, HAU, and PT-F. Datasets are available for shoot DW yields, shoot metal concentrations and shoot metal removals. The influences of soil contamination levels, fertilization, maintenance through compost dressing, plant species and genotypes (mother-line, somaclonal tobacco variants, and sunflower mutants), agricultural practices such irrigation, co-cropping and flower topping were considered. Touro (ES): Tobacco: Shoot DW yield was primarily dependent on the climatic conditions during the growth season and varied accordingly: the highest biomass for all genotypes was achieved in the 2014 harvest and reached 3400-4000 kg DW ha-1 (levels comparable to those obtained in the agricultural soils of Bettwiesen and Lommel). Differences between the BAG motherline and the 10-8 and 10-4 variants were not pronounced, and biomass tended to be higher for BAG. Shoot Cu removal in 2014 (60-70 g Cu/ha) was lower than that obtained in Biogeco. Piedrafita (ES): At this site sunflower could only grow in the compost-amended plots, while tobacco could grow in both compost-amended and untreated mine-soils (after fertilization with inorganic NPK). Tobacco: Annual shoot DW yield again varied widely according to climatic conditions and also to competition with weeds: biomass production was highest in 2012. There were no consistent differences between motherlines and other variants. Biomass production and Cd/Zn extraction potential was significantly lower than that observed in Lommel and Bettwiesen. Sunflower: Sunflower cultivation was more successful than tobacco at this site. Annual shoot DW yields were similar to that obtained at the other Greenland field sites. Cd/Zn extraction potential were similar to that obtained in Bettwiesen and Lommel in the 2012 and 2013 harvests, but lower in 2014. Mutant 1 reached up to 6772 and 23 g/ha Zn and Cd extraction potential, respectively. Biogeco (FR): Tobacco: its shoot DW yield depended on total soil Cu, soil amendments, and the genotype in some plots. Flower topping in years 6 and 7 allowed the development of bottom suckers, which increased the shoot biomass. Depending on climatic conditions, early flower topping in Southwest France allowed to harvest tobacco shoots two to three times per year and to avoid loss of dried leaves. ● At moderate soil Cu contamination (258 – 382 mg Cu/kg): in year 6, shoot Cu removal by the OMDL plants reached 84-132 g Cu/ha, without significant genotype influence. The second compost dressing in year 6 (OM2DL) reduced shoot Cu removal as compared to OMDL, likely due to a decrease in mobile soil Cu. In year 7, the OM2DL plants had a higher shoot yield than the OMDL ones. ● At high soil Cu contamination (894 – 1020 mg Cu/kg): in year 6, the 10-8 variant best performed in only one OMDL plot thanks to its higher biomass. Shoot Cu removals varied from 68 to 193 g Cu/ha. For the OM2DL plots, the tobacco genotype did not significantly influence shoot Cu removal and its values were similar or higher. In year 7, shoot length and DW yield were higher for the OM2DL plants than for the OMDL ones, which promoted shoot Cu removal. This was sometime more marked for the 10-8 variant. ● For the Linz-Donawitz slag (LDS) amended-plots in year 6, shoot Cu removal peaked with the variant 10-8 in one plot, reaching 254 g Cu/ha. For all genotypes, differences were not significant between the P-LDS and Unt plots. The LDS plants produced less shoot DW yield than the OM2DL ones. In year 7, tobacco from untreated and LDS-amended plots had again a lower stem length than the OM2DL plants. Differences between genotypes were significant only in 3 plots, 10-8 variant best developing in two cases, but no general trend was found. The influence of soil amendment, especially the long-term effect of the compost second dressing was key factor to explain the shoot yield. Sunflower: At Biogeco, in year 6 and 7, leaf chlorosis occurred on many sunflower plants growing in OMDL amended plots with high total soil Cu and in both LDS-treated soils. No sunflowers were growing in the UNT soil. ● at moderate soil Cu contamination (258 – 382 mg Cu/kg): in year 6, the M2 mutant performed best in comparison to other genotypes, without effect of the second compost dressing. The shoot DW yield of the M3 mutants was lower than that of motherline (IBL04) plants in the OM2DL plots. Overall, shoot Cu concentrations were similar for all plants. The M2 mutant showed a higher shoot Cu removal than the other genotypes in the OM2DL plot. Shoot Cu removal was in the 42 g Cu/ha range. Due to lower shoot Cu concentrations than previous years, this was lower than the 100 g Cu/ha reported in years 1 and 2. In year 7, shoot DW yield varied from 0.5 to 36 t/ha depending on plots and genotypes. ● at high soil Cu contamination (894 – 1020 mg Cu/kg): Sunflower plants best developed in years 6 and 7 on all OM2DL plots that got a second compost dressing in year 6, which illustrated the necessity to maintain soil fertility through organic matter. For the genotype influence, the M2 mutant displayed higher stem length in some plots. A single compost dressing increased the shoot DW yield more than the incorporation of LD slags. The Carmeuse-LDS

13 Appendix to WP1 scientific report – Greenland Final Report

and P-LDS had a similar influence on sunflower growth, and both M3 and IBL04 plants died in these plots during summer. In year 6 for the OM2DL plots, M2 mutants produced a higher shoot biomass than other genotypes. The same trend was noticed in the OM amended-plot but the M2 variant only differed from other variants. Other genotypes had similar shoot DW yield in the OM2DL plots. Both mother-line and M3 mutant developed only in the plots with recent compost dressing (OM2DL and OM). Shoot Cu concentration was the highest in the OMDL plots for the M3 variants followed by the M2 ones. For the M3 plants, this corresponded with their low shoot DW yield. All mutants showed the lowest shoot Cu concentrations in the OM2DL plots compared to the OMDL plots demonstrating the influence of the second compost dressing. The M1 mutant contained a higher shoot Cu concentration when grown on the OM plot compared to other plants. The low shoot Cu concentrations (close to the upper critical threshold value for Cu in higher plants) for the OM2DL plants corresponded with their high shoot biomass, suggesting a dilution effect. Shoot Cu removal was the highest for the M2 mutant in the OM2DL plots, thanks to their high shoot biomass. The trend was similar for M2 mutants from the OM plots, notably compared with the IBL04 plants. In overall, the high shoot biomass would be the main driver for the high shoot Cu removal. According to the plant density, plots and sunflower genotypes, shoot Cu removal varied in the 21-105 g Cu/ha range. ● LDS plots: only M1 and M2 mutants developed on these plots with a better growth of M1 plants. Shoot Cu concentrations of M1 and M2 mutants were higher in both LDS amended plots than in the OM2DL plots. High values for shoot Cu removal were obtained for the M1 mutant in both LDS plots, mostly due to their higher biomass. In year 7 the shoot DW yield was enhanced in the OM2DL plots, and the LDS amended plots. Sunflowers were unable to grow on the untreated plots. The influence of the genotype was insignificant, but the second compost dressing in year 6 was the key factor to promote the shoot yield and Cu removal. Between years 4 and 6, extractable Cu fraction in the OMDL plots was reduced by 38%. Since the experiment started, shoot Cu removal fit with a quadratic function, likely following reactions of compost with Cu, nutrient release from compost decay and bioavailable Cu stripping. Lommel (BE): phytoextraction of Cd/Zn was assessed. Tobacco clones and sunflower mutants were cultivated from years 1 to 4. Brassica napus and Cannabis sativa (hemp) were also implemented. In 2013, shoot DW yield of tobacco was similar for all genotypes ranging from 1.3 to 1.7 t/ha. For sunflower, shoot DW yield varied from 3.5 to 7.5 t/ha, with the M1 mutant producing lower shoot biomass than the control plants and the other mutants. Hemp developed well and its shoot DW yield reached 17.5 t ha-1 yr-1. Shoot metal removals for tobacco were in the following ranges (in g/ha/yr): Cd 14 – 20, Pb 8.6-11.9, Zn 252-331. Genotype had an influence (e.g. higher shoot removal for Cd/10-8 variant and Zn/7-19 variant). Compared to tobacco, shoot metal removals of sunflower were higher for Zn (1992-2504) and slightly higher for Cd (17.7-23.6) and Pb (17.7-42.5), with a genotype influence. The phytoextraction by hemp was 7 g Cd, 41 g Pb, and 1355 g Zn/ha. The phytoextraction of Cd, Pb and Zn by all tested sunflowers was higher than that of the tobacco clones and hemp. Tobacco clones had higher shoot Cd and Pb concentrations but the higher shoot DW yield of sunflowers lead to higher shoot metal removals. Hemp production on metal-contaminated soil could be relevant if cutting the pollutant linkage to food along with an economic profit from the plant-based feedstock (e.g. fibre) is the primary goal instead of the other ecosystem services (e.g. decreases of labile metal pools, decontamination, and soil remediation). In 2014, the biomass production, and consequently the shoot metal removal, was higher than in 2013, especially for tobacco (tobacco 3.6-4.9 t/ha; sunflower 5.8-9.5 t/ha). Shoot removals were 59.6-122 g Cd, 38-70 g Pb and 1027-1926 g Zn/ha for tobacco, 32-61 g Cd, 17.5-34.5 g Pb, and 2624-5745 g Zn/ha. This highlights the influence of annual climatic conditions on shoot metal removals. Bettwiesen (CH): The crop rotation is based on four sunflower and five tobacco genotypes with higher metal tolerance and accumulation properties for stripping bioavailable Zn and Cd excess in topsoil. After 5 years, the labile Zn pool in soil was lowered by 45-70%, and up to 67% for Cd and 62% for Pb. A Mass Balance Analysis confirmed soil Zn decontamination in line with plant Zn uptake. The plants partially take Zn from the non-labile pool of the total. Moreover the results confirm a strong immobilization effect of the plant rhizosphere (by increasing soil pH up to one unit) due to phytoextraction treatment. The ‘stripping’ of bioavailable Zn is feasible within a few years period. To decrease available soil Zn below the Swiss threshold value, the phytomanagement would take 3-12 years at moderate available Zn levels and 5-25 at high levels. Various plant densities and intercropping of sunflower and tobacco with fodder pea during winter until in early spring, and wild type Galingsoga parviflora are further explored.

14 Appendix to WP1 scientific report – Greenland Final Report

Due to long lasting cold and rainy weather condition in 2014, the biomass production, and consequently the shoot metal removal on the B3 and B4 plots was lower than in 2011-13, especially for tobacco (tobacco 1.9-4.1 t/ha; sunflower 3.5-11.7 t/ha). Shoot tobacco removals were 381-1148 g Zn/ha, 3-11 g Cd/ha, 25-80 g Cu/ha, and for the sunflower 815-3528 g Zn/ha, 1-14 g Cd/ha, 83-193 g Cu/ha. This highlights the influence of annual climatic conditions on shoot metal removals. The metal extraction efficiency of the tobacco and sunflower mutants and in vitro bred cultivars and its controls (motherline inside the brackets) followed the following hierarchy for the experimental years 2011-14: tobaccos: NBCu10-8 > NBCu10-4 > (BAG) sunflowers: mutant 3 ≅ mutant 1 > mutant 2 > (IBL04). Mutant 4 additionally tested in 2014 was best for shoot yield, Cd and Cu extraction. In spite of good shoot yield of the tobacco and sunflower controls, their shoot metal extraction efficiency was constantly lower, compared to the sunflower mutants and in vitro optimized variants of tobacco. Based on representative and randomly taken soil samples on the B3 and B4 plots, the phytoextraction efficiency along the four years period (2011-14), was 73–94% for labile Zinc and 73-95% for labile Cd topsoil concentrations. 1.3.3.2. (aided) phytoextraction using woody SRC The capacity of poplar and willow to colonize hostile environments such as mine wastes is recognized. Numerous Salix and Populus clones have been screened, and show great variation in biomass production, TE tolerance and accumulation patterns in roots and leaves between clones. Most promising poplars and willows (according to climatic conditions and shoot TE concentrations) were assessed at 6 sites: Lommel (HAU), Högbytorp and French trial (SLU), Freiberg (LfULG), Piedrafita (ES) and Phytagglo (FR). Freiberg, Halsbrücke Krummenhennersdorf (DE): this 9-old field trial is a SRC plantation on contaminated agricultural land. Shoot DW yields reached 15 t/ha/yr for poplar SRC and 14-19 t/ha/yr for willow SRC, corresponding to common values (6-16 t DM ha-1 yr-1). Stem wood and bark ionomes of poplars and willows depended on genotypes, particularly for Cd (3-10 mg/kg in wood; 8-30 mg/kg in leaves), and clonal differences were higher across willows. Values were higher in willows compared to poplars. Willow cultivars Tora, Tordis and Gudrun displayed the highest wood and foliar Cd concentrations among all cultivated clones of poplars and willows. The ratio of the 3-years-average concentration in leaves compared to wood was about 2.6 and 2.7 for Cd and about 5.7 and 6.2 for Zn in poplar and willow, respectively. Bark Cd concentrations account for triple (poplars) to fourfold (willows) of those in stem wood while Pb-concentrations did not differ between the compartments. Foliar As-concentrations were mostly below the detection limit. Wood As concentrations varied from 2 to 4 mg/kg. For the third rotation, willow Tora produced the highest biomass out of all poplars and willows (followed by Tordis) and it displayed relatively high Cd and Pb accumulation capacity. This confirmed Tora as a relevant choice for metal (Zn, Cd) phytoextraction, and Tordis as well (for Cd). Piedrafita do Cebreiro (ES): plots with Salix smithiana and Salix atrocinerea, in monoculture and inter-cropped with Alnus glutinosa, and Salix cv. Tora were established in autumn 2012. Plots were unamended or amended with 5% (w/w) compost. Plant survival was significantly higher for S. smithiana than either S. atrocinerea or Tora. Plant growth (height/spread) and leaf/stem Cd/Zn concentrations were recorded in 2013 and 2014. No significant effects of inter-cropping on plant growth have been recorded to date (although there was a tendency to increased tree height in intercropped plots in 2012). Intercropped S. smithiana plots showed significantly higher leaf Cd and Zn concentrations (approx.. 12 and 1400 mg/kg). Significant changes in total or NH4Cl-extractable concentrations of Cd and Zn have not yet been observed. Changes in soil exposome: At Freiberg, generally total soil Cd decreased (e.g. 3.4 to 2.9 mg/kg for Weser 6 poplar; 2.5 to 1.0 for Tora willow) between 2011 (year 6) to 2013 (year 8). The rhizosphere of willow and poplar clones at contaminated and adjacent reference sites showed a lower pH for SRC compared to arable land (with a higher decrease for willow SRC). Since the SRC implementation in 2005, the initial soil pH of 5.7 (CaCl2) dropped in average to 5.2. However, soil pH increased in poplar plots after the harvest in year 8, whereas a further decrease occurred for willow plots especially for Tordis cultivar. NH4NO3-extractable soil Cd and Pb were roughly 10 fold higher under SRC and grassland compared to annual cropped land. In contrast, mobile As in soil decreased for SRC (3 to 9 fold) compared to annual cropped land (lowest values for willow SRC). Soil pH was higher under winter wheat with the lowest NH4NO3-extractable soil Cd in year 2. NH4NO3-extractable soil Cd decreased for Jorr, Sven, Tora willow SRC and Max 3 poplar SRC.

15 Appendix to WP1 scientific report – Greenland Final Report

Soil microbial communities: At Freiberg, arylesterase and arylsulfatase activities did not differ between SRC and control plots. Alkaline phosphatase was higher in SRC plot. At Piedrafita do Cebreiro, shifts in the structure of the bacterial communities, i.e. total Eubacterial community, Alpha- and Beta-proteobacteria, Actinobacteria and Streptomycetaceae, were assessed over two years in the compost-amended plots (aided-phytoextraction) and unamended plots (phytoextraction), the vegetation cover being SRC (with Salix sp.). The DGGE profiles of soil samples collected before GRO implementation (time=0) are always separated from phytomanaged soils (similarity <20%). In general, compost-amended plots were separated from untreated soils, although this separation was more pronounced in year 2. In both years 1 and 2, there are some clear separations between soil groups under the different vegetation covers, and between mono- and co-cultivated Salix smithiana. The Shannon diversity index indicated a decrease in the diversity of these bacterial communities with time (from 1 to 2 years). A higher diversity always occurred in phytomanaged soils compared to untreated soils, without clear plant-induced effect. After 3 years of SRC, most soil enzyme activities increased in the phytomanaged soils, with a plant-induced effect, and this was more pronounced in the compost-amended soils than the unamended soils. However, a clear effect of the co-cropping of Salix smithiana with Alnus glutinosa was less evident. At Phytagglo (FR), 350 willows (Salix viminalis) were planted on 9 raws in April 2013. Inter-rows were covered with beech mulch. This alkaline technosol (pH 8.1) developed on dredged sediments is contaminated by Zn (1117), Pb (262), Cu (100) and Cd (2.9), with a high spatial variability for Cd and Zn and a relatively high OM content (2.3%). In some plots, citric acid based product, ferrous sulfate, and elemental S were separately incorporated in order to investigate the effects of soil acidification. NH4N03-extractable metal fractions were low (e.g. 0.014% for Cd, 0.2% for Zn). In year 1, extractable metal fraction decreased for Pb, and remained steady for Cu, Cd and Zn. The survival rate of willows was 90% in year 1. Their foliar Cd and Zn concentrations ranged between 1.5-3.8 and 233-1176 mg kg-1, which is quite high when considering the low Cd and Zn mobility and alkaline soil pH. Foliar Cu concentrations (7-12 mg kg-1) were similar to common values of willows on uncontaminated soils. In year 1, maximum stem height increased by 4 fold and trunk diameter by 2 fold. After one month, soil pH did not decrease in the amended soils and further inputs of potentially acidifying compounds are planned. Högbytorp (landfill leachate trial, SE): For 14 commercial willow SRC plantations long-term grown (ca. 15 years) on agricultural soil in Sweden, total topsoil Cd decreased (ca. 13% on average) compared to adjacent fields cultivated with cereals in common crop-rotations. The biomass productions on these SRC fields were lower than the indicative 10 t DM ha-1 yr-1 expected nowadays in well-managed fields. Farmers had lack of experience in growing such crops, and beneficial incentives in terms of subsidies caused limited engagement throughout the process. Here, treatments consist of three supply rates of landfill leachate (irrigation started in 2005 and carried out until 2010) and a control, with two willow clones, i.e. Tora (Salix schwerinii x viminalis) and Gudrun (Salix dasyclados variety with partly Russian origin, more frost-tolerant than Tora). For willows, shoot concentrations varied in the 1-4.5 mg Cd and 40-120 mg Zn kg-1 DW ranges. Tora showed higher shoot Cd, Co, Mn, Pb and Zn concentrations than Gudrun for leachate irrigated-plots. Tora had higher shoot Cd, Co and Zn concentrations for the plots irrigated at the second supply rate compared to other treatments and the control. Gudrun displayed higher shoot concentrations for all metals in the control plots. The leachate treatments did not influence the shoot Cr and Cu concentrations in both clones, even though these concentrations were low for Tora on the control plots compared to the treatments 1-3. Nickel was the only metal with higher shoot concentrations in Gudrun for all treatments. Total shoot N concentration was roughly similar for both clones, except a higher concentration for Gudrun on the control plots. Lommel (BE): Willow clones ‘Belgisch Rood (BR)’ (Salix x rubens var. basfordiana) and ‘Tora’ (Salix schwerinii x Salix viminalis) were compared. Shoot DY yields after 3 years were 5.4 (BR) and 9.0 (Tora) t/ha. Again shoot Cd concentrations were higher in Tora (30 mg/kg) than in BR (24 mg/kg) and shoot Zn concentration as well (1268 and 918 mg Zn/kg). The Tora willow clone had a Cd and Zn removal capacity (274 g Cd and 11 417g Zn ha-1) which is twice as high as that for BR clones. Both, the higher biomass production (ton ha-1) and metal uptake capacity (mg kg-

1 DW) make the TO willow clone the favorable clone to select for phytoextraction applications. 1.3.3.2. TE hyperaccumulators: can accumulate high concentrations of metal(loid)s (e.g. Cd, Ni, Zn, Se, and As) in their above-ground biomass and possess some economic added value (renewable biomass for bio-economy and bio-ores). Variations in both biomass production and TE accumulation within populations of hyperaccumulators, such

16 Appendix to WP1 scientific report – Greenland Final Report

as Noccaea caerulescens (for Zn/ecocatalysis), Alyssum murale, A. bertolonii and A. corsicum, (for Ni phytomining) allows for the selection and breeding of improved phytoextractor plants. The main bottleneck limiting their practical application is the low biomass production of most species (except some Ni-hyperaccumulators) and the number of cropping cycles required for clean-up. However this number is generally reduced when the option of bioavailable TE stripping is considered. Additional limiting factors include the absence of commercially available seeds/seedlings, their sensitivity to the presence of contaminants other than the hyperaccumulated TE, a general lack of knowledge related to their cultivation, climate needs and competition with other TE-tolerant plants.

Piedrafita (ES): sub-plots were established in May 2012 with the Cd/Zn hyperaccumulator Noccaea caerulescens and inter-croppings of N. caerulescens with Lupinus albus and Lotus corniculata. Unamended plots were fertilized with inorganic commercial fertilizers. Nitrogen fixation can decrease soil pH due to nitric acid accumulation in the rhizosphere, which for phytoextraction purposes can in turn induce an increase in TE bioavailability to the co-cropped TE-accumulators. Inter-cropping the Cd/Zn-hyperaccumulator N. caerulescens with the legume Lotus corniculatus tended to increase Cd accumulation by the hyperaccumulator. Other Lotus species show potential for incorporating into GRO due to their worldwide distribution and high adaptation to a number of abiotic stresses. Candidates with good potential for cultivation in degraded or marginal soils include Lotus corniculatus, L. uliginosus, L. tenuis and L. creticus. Restoration of soil microbial activities and community: The DGGE profiles showed similar clusters for the soils planted with N. caerulescens and those co-cropped with L. corniculatus. The separation between unamended and compost-amended soils was clearer in year 2. Unplanted soils formed clusters, and were well separated from planted soils. Mono-cultures and co-cultures tended to cluster separately. The diversity of bacterial communities increased with time, with a higher diversity in phytomanaged soils compared to untreated soils, and a higher diversity in planted soils compared to unplanted soils. Phytagglo (FR): Seven plots were set up with potentially acidifying properties, i.e. legume plant (Lupinus albus), a citric acid based product, peat-like, ferrous sulfate and elemental sulphur, Arabidopsis halleri and control. Reppel (BE/FR): Since 2004, Pteris vittata L., an As hyperaccumulator was cultivated for bioavailable As stripping in this Belgian soil polluted by atmospheric fallout. Generally, frond DW yield was doubled in the contaminated soils compared to the uncontaminated control soil. Soil treatments, i.e. Beringite (B, 5% w/w), iron grit (Z, 1% w/w) and their combination (BZ), and season did not influence annual frond yield, except differences between B and BZ in November and between November and May for the untreated (Unt) and B soils. On the 2006-2013 period, leachate As concentration remained lower in Z-treated soils than in the Unt and B soils. Mean values of frond As concentrations (in mg As/kg) varied in the 60-171 range for the control soil and in the 970-2870 range for the contaminated soils. Frond As removal varied from 3.89 to 2.28 g As/m² in the decreasing order: Unt, B > BZ, Z. For vegetables cultivated after several fern crops, root DW yield of lettuce was higher in BZ-treated soils and lower in the B-treated ones, whereas shoot DW yield did not differ across soil treatments. Shoot As concentration (in µg/g DW) varied from 1.28±0.25 to 2.5±0.5 and was lower in BZ-, B-, and Z-lettuces.